Transformation of acyclic alkenes to hydrido carbynes by (PNP R)Re complexes

Published on Web 04/30/2004

Transformation of Acyclic Alkenes to Hydrido Carbynes by

(PNP^R)Re Complexes

Oleg V. Ozerov,^†Lori A. Watson, Maren Pink, and Kenneth G. Caulton*

Contribution from the Department of Chemistry, Indiana UniVersity,

Bloomington, Indiana 47405

Received December 10, 2003; E-mail: caulton@indiana.edu

i

t

Abstract: Synthesis of (PNP^R)ReOCl₂(PNP^R) (R₂PCH₂SiMe₂)₂N, R ) Pr, Cy, and Bu) from (Me₂S)₂-

ReOCl₃and (PNP^R)MgCl is described. Magnesium and H₂convert (PNP^R)ReOCl₂first to (PNP^R)ReO(H)₂

and then to (PNP^R)Re(H)₄, the last being an operationally unsaturated species which can bind PMe₃or

p-toluidine. Acyclic alkenes react with (PNP^R)Re(H)₄at 22 °C to give first (PNP^R)Re(H)₂(olefin) and then

(PNP^R)ReH(carbyne), in equilibrium with its η²-olefin adduct. Re can also migrate to the terminal carbon of

internal olefins to form a carbyne complex. Allylic C-SiMe₃or C-NH₂bonds are not broken, but OEt,

OPh, and F vinyl substituents (X) are ultimately cleaved from carbon to give the RetC-CH₃complex and

liberate HX. DFT calculations, together with detection of intermediates for certain olefins, help to define a

mechanism for these conversions.

Introduction

We reasoned that a metal complex which was more π-basic

The chemistry of the fragment RuHClL₂(L ) PⁱPr₃), which

exists as the chloride-bridged dimer¹[RuH(µ-Cl)L₂]₂but gives

monomeric products with Lewis bases and olefins, is satisfac-

torily rationalized as being a π-electron rich species “in search

of” π-acid ligands. Of special interest, it has the kinetic and

thermodynamic ability to isomerize certain olefins to carbene

and even carbyne ligands.^1-13This characteristic reactivity was

attributed to the absence of electron-withdrawing π-acid ligands

in RuHClL₂. However, it was found that, to be isomerizable,

an olefin needs a π-donor substituent (OR, NR₂, F). This is

then a thermodynamic demand, because even the free carbene

:C(X)R is stabilized by π-donor substituent X. This explains

why hydrocarbon olefins are not isomerized to carbene com-

plexes by RuHClL₂. Related transformations of unsaturated

hydrocarbons have involved iridium,^14-17osmium,¹⁸and ru-

thenium.¹⁹

than RuHClL₂might convert olefins to an isomeric form which

is more π-acidic. Our approach to such a complex was to go to

a 5d metal (generally more “reducing” than a 4d metal), in a

lower oxidation state than Ru(II). Our choice was rhenium,

which already has some history as a source of enhanced reducing

power.^20,21In addition, there is evidence^22-24that an amide, NR₂,

is more π-donating than chloride, and thus we sought a rhenium

complex with an amide ligand. Since amides are prone to â-H

migration to make hydrido imine complexes,^25-29it was

attractive to have silicon as the substituent on N. This led us to

attempt to install ligands of the sort pioneered by the Fryzuk

group:^30-36

(14) Crabtree, R. H.; Milhelcic, J. M.; Quirk, J. M. J. Am. Chem. Soc. 1979,

101, 7738.

(15) Kanzelberger, M. Singh, B.; Czerw, M.; Krogh-Jespersen, K.; Goldman,

A. S. J. Am. Chem. Soc. 2000, 122, 11017.

(16) Krogh-Jesperson, K.; Czerw, M.; Summa, N.; Renkema, K. B.; Achord,

P. D.; Goldman, A. S. J. Am. Chem. Soc. 2002, 124, 11404.

(17) Renkema, K. B.; Kissin, Y. V.; Goldman, A. S. J. Am. Chem. Soc. 2003,

125, 7770.

(18) Edwards, A. J.; Esteruelas, M. A.; Lahoz, F. J.; Lopez, A. M.; Onate, E.;

Oro, L. A.; Tolosa, J. I. Organometallics 1997, 16, 1316.

(19) Borowski, A. F. Sabo-Etienne, S.; Christ, M. L.; Donnadieu, B.; Chaudret,

B. Organometallics 1996, 15, 1427.

(20) Leeaphon, M.; Ondracek, A. L.; Thomas, R. J.; Fanwick, P. E.; Walton,

R. A. J. Am. Chem. Soc. 1995, 117, 9715.

(21) Spaltenstein, E.; Erikson, T. K. G.; Critchlow, S. C.; Mayer, J. M. J. Am.

Chem. Soc. 1989, 111, 617.

(22) Fryzuk, M. D.; MacNeil, P. A.; Rettig, S. J. Organometallics 1986, 5, 2469.

(23) Collman, J. P.; Sears, C. T. Inorg. Chem. 1968, 7, 27.

(24) Rahim, M.; Ahmed, K. J. Inorg. Chem. 1994, 33, 3003.

(25) Mayer, J. M.; Curtis, C. J.; Bercaw, J. E. J. Am. Chem. Soc. 1983, 105,

2651.

(26) Tsai, Y.-C.; Johnson, M. J. A.; Mindiola, D. J.; Cummins, C. C. J. Am.

Chem. Soc. 1999, 121, 10426.

^†Present address: Department of Chemistry, Brandeis University,

MS015, Waltham, MA 02454.

(1) Coalter, J. N.; Huffman, J. C.; Streib, W. E.; Caulton, K. G. Inorg. Chem.

2000, 39, 3757.

(2) Ferrando, G.; Coalter, J. N.; Gerard, H.; Huang, D.; Eisenstein, O.; Caulton,

K. G. New J. Chem. 2003, 27, 1451.

(3) Ferrando-Miguel, G.; Gerard, H.; Eisenstein, O.; Caulton, K. G. Inorg.

Chem. 2002, 41, 6440.

(4) Ferrando-Miguel, G.; Coalter, J. N.; Gerard, H.; Huffman, J. C.; Eisenstein,

O.; Caulton, K. G. New J. Chem. 2002, 26, 687.

(5) Ferrando, G.; Gerard, H.; Spivak, G. J.; Coalter, J. N.; Huffman, J. C.;

Eisenstein, O.; Caulton, K. G. Inorg. Chem. 2001, 40, 6610.

(6) Coalter, J. N.; Huffman, J. C.; Caulton, K. G. Chem. Commun. 2001, 1158.

(7) Coalter, J. N.; Caulton, K. G. New. J. Chem. 2001, 25, 679.

(8) Caulton, K. G. J. Organomet. Chem. 2001, 617, 56.

(9) Coalter, J. N.; Bollinger, J. C.; Eisenstein, O.; Caulton, K. G. New J. Chem.

2000, 24, 925.

(10) Coalter, J. N.; Ferrando, G.; Caulton, K. G. New J. Chem. 2000, 24, 835.

(11) Coalter, J. N.; Huffman, J. C.; Caulton, K. G. Organometallics 2000, 19,

3569.

(12) Coalter, J. N., III; Streib, W. E.; Caulton, K. G. Inorg. Chem. 2000, 39,

3749.

(13) Coalter, J. N.; Bollinger, J. C.; Huffman, J. C.; Werner-Zwanziger, U.;

Caulton, K. G.; Davidson, E. R.; Gerard, H.; Clot, E.; Eisenstein, O. New

J. Chem. 2000, 24, 9.

(27) Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7010.

(28) Bryndza, H. E.; Tam, W. Chem. ReV. 1988, 88, 7010.

(29) Fryzuk, M. D. M., C. D. Coord. Chem. ReV. 1989, 95, 1.

(30) Fryzuk, M. D.; Giesbrecht, G. R.; Rettig, S. J. Can. J. Chem. 2000, 78,

1003.

(31) Fryzuk, M. D.; Leznoff, D. B.; Rettig, S. J.; Young, V. G., Jr. J. Chem.

Soc., Dalton Trans. 1999, 147.

9

J. AM. CHEM. SOC. 2004, 126, 6363-6378

6363

A R T I C L E S

Ozerov et al.

Scheme 2

employed by Wilkinson in the synthesis of (R₄N)₂ReOCl₅.³⁸

The material was isolated in good yield and worked superbly

in our syntheses of (PNP^R)ReOCl₂(Scheme 3), although it may

We communicated earlier on synthetic access to (PNP^R)-

Re(H)₄and its reaction with ethylene.³⁷We now report on the

breadth and scope of this reaction with terminal and internal

acyclic olefins, together with density functional theory (DFT)

calculations designed to guide mechanistic analysis of these

unusual and facile reactions. They indeed transform olefins in

ways which are impossible for RuHClL₂.

Scheme 3

Results

Installing the PNP^RLigand on Re. We have previously

reported³⁷the synthesis of (PNP^Cy)ReOCl₂and its reduction

under hydrogen atmosphere to (PNP^Cy)Re(H)₄. Here we also

report the preparation of analogous PNP^iPrand PNP^tBucom-

pounds and the reactivity of these tetrahydrides with olefins.

The preparation of the pincer ligands as their lithium salts was

performed according to our modification of the original

procedure by Fryzuk. These (PNP^R)Li (1) derivatives were

isolated and then converted to (PNP^R)MgCl (2). We have

previously reported³⁷that the usage of PNP^Cymagnesium

derivatives in the syntheses of (PNP^Cy)ReOCl₂resulted in

superior yields compared to their lithium counterparts. (PNP^tBu)-

MgCl (2c) can be conveniently isolated as its 1,4-dioxane

adduct, but (PNPⁱPr)MgCl (2b) can only be obtained as viscous

oil and was therefore prepared in situ and used without isolation

(Scheme 1).

contain a small amount of iodide (see Experimental Section).

We also prepared the bromide analogue of 3, yellow-green 3-Br

(Scheme 2). 3-Br served well as a Re starting material in the

preparation of (PNP^Cy)ReOBr₂.

The new Re complexes (PNP^iPr)ReOCl₂and (PNP^tBu)ReOCl₂

have similar properties to the previously reported (PNP^Cy)-

ReOCl₂. All three green compounds exist as mixtures of two

isomers and are highly soluble in hydrocarbon solvents. Our

previous report³⁹detailed the thermal isomerization of (PNP^Cy)-

ReOCl₂(4a) into (POP^Cy)ReNCl₂(6a). The new complexes

(PNP^iPr)ReOCl₂(4b) and (PNP^tBu)ReOCl₂(4c) undergo a similar

transformation to form 6b and 6c, respectively. For the latter,

this isomerization occurs at substantially lower temperatures.

In fact, a small amount (5-10%) of (POP^tBu)ReNCl₂(6c) is

formed even during the time of synthesis and isolation of

(PNP^tBu)ReOCl₂(1-2 h at 22 °C). Solid samples of (4c) were

stored in the freezer at -30 °C for months without change. For

4b, the rate of this isomerization is approximately the same as

for 4a.

Installation of the Hydride Ligands. The tetrahydrides 5a-c

were prepared by Mg reduction of 4a-c under H₂atmosphere

(Scheme 3). 5b and 5c are considerably more lipophilic than

5a and therefore could only be recrystallized in good yield from

solvents that, in our experience, are less solubilizing than

pentane such as neo-hexane, tetramethylsilane, and hexameth-

yldisiloxane.

Scheme 1

We were unsatisfied with the rather lengthy purification of

(PNP^Cy)ReOCl₂synthesized from (Ph₃PO)(Me₂S)ReOCl₃. The

preparation of compounds 4 can be substantially simplified by

using (Me₂S)₂ReOCl₃(3) as a starting material, which generates

only inorganic salts (MgCl₂) or volatile Me₂S upon reaction

with 2. Additionally, its poor solubility, while not significantly

retarding reactivity, permits easy separation of unreacted 3.

(Me₂S)₂ReOCl₃was prepared via reduction of an HCl solution

of HReO₄by NaI in the presence of Me₂S (Scheme 2).

Reduction of Re(VII) to Re(V) by iodide anion has been

The NMR spectra of 5b-c display the same key features as

those of 5a. The overall symmetry in solution is C_2V, and a single

triplet of 4H intensity is observed at high field (t, J_H-P) 22

Hz, at -9.42 ppm for 5b). At the same time, the aliphatic region

(32) Fryzuk, M. D.; Leznoff, D. B.; Thompson, R. C.; Rettig, S. J. J. Am. Chem.

Soc. 1998, 120, 10126.

(33) Fryzuk, M. D.; Leznoff, D. B.; Ma, E. S. F.; Rettig, S. J.; Young, V. G.,

Jr. Organometallics 1998, 17, 2313.

(34) Fryzuk, M. D.; Gao, X.; Rettig, S. J. Organometallics 1995, 14, 4236.

(35) Fryzuk, M. D.; Gao, X.; Rettig, S. J. J. Am. Chem. Soc. 1995, 117, 3106.

(36) Fryzuk, M. D. Montgomery, C. D.; Rettig, S. J. Organometallics 1991,

10, 467.

(37) Ozerov, O. V.; Huffman, J. C.; Watson, L. A.; Caulton, K. G. Organo-

metallics 2003, 22, 2539.

1

of the H NMR spectra of 5b-c is considerably more useful

than that of 5a, for which it is essentially uninterpretable because

(38) Grove, D. E.; Wilkinson, G. J. Chem. Soc. A 1966, 1224.

(39) Ozerov, O. V.; Gerard, H. F.; Watson, L. A.; Huffman, J. C.; Caulton, K.

G. Inorg. Chem. 2002, 41, 5615.

9

6364 J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004

Transformation of Acyclic Alkenes to Hydrido Carbynes

A R T I C L E S

Scheme 5

of the multiple peaks arising from the Cy groups. Thus, 5b

shows a single methine signal and two signals belonging to the

diastereotopic methyls of the isopropyl group in addition to one

singlet (12H) arising from the Si-Me groups and one virtual

triplet (4H) from the SiCH₂P methylenes.

The reduction of 4a-c by Mg must proceed in more than

one step. NMR analysis of the reaction mixtures derived from

4a at intermediate reaction times revealed (Scheme 4) the

Scheme 4

broadening of the hydride ¹H NMR signal and of the ³¹P NMR

signal is observed upon exposure to H₂, indicative of the

reversible H₂adduct formation. The analysis of the possible

isomers of “(PNP^Cy)ReH₆” (8a) by means of DFT calculations

and by NMR has been reported.³⁷In contrast, no change in the

NMR spectrum of a C₆D₆solution of 5c is observed upon

addition of 4-picoline. Evidently, the PNP^tBuligand is too bulky

to permit coordination of another ligand to Re in 5c. Addition

of 1 equiv of 4-picoline to C₆D₆solutions of 5a or 5b causes

broadening of the NMR resonances of both 5a(b) and 4-picoline,

as well as lightening of the purple-red color of 5. Removal of

the volatiles from these solutions regenerates pure 5a-b

quantitatively. The reaction of 5a with 1 equiv of PMe₃, on the

other hand, results in ca. 99% conversion to the colorless adduct

(PNP^Cy)ReH₄(PMe₃) (10a), and exposure to vacuum results in

the loss of H₂(and not PMe₃) and formation of the blue-purple

(PNP^Cy)Re(H)₂(PMe₃) (11a). Reactions with H₂, 4-picoline, and

PMe₃demonstrate the adaptability of the amido ligand which

allows fast and reversible substrate binding.

presence of (PNP^Cy)ReO(H)₂(7a), which completely disappears

by the end of the reaction. Apparently, the removal of Cl ligands

from Re occurs before the removal of the oxo group. The yellow

complex 7a was independently synthesized from 4a and 2 equiv

of NaBEt₃H (Scheme 4). The NMR spectra are consistent with

an average C_2Vsymmetry in solution, and the hydridic signal

(2H) is observed as a triplet at 5.59 ppm. Such a downfield

shift for a non-d⁰metal hydride is atypical, but in several

reported complexes that contain hydrides cis to an oxo group,

1

the hydrides resonate in the downfield region of the H NMR

spectrum. On the basis of this spectroscopic evidence, we

propose the mer,trans structure for this compound. NMR

resonances attributable to 7b and 7c were also observed at

intermediate stages of the Mg/H₂reductions of 4b and 4c. In

contrast to 4, 7a is thermally stable (no change after 24 h at 90

°C in C₆D₆) and only a single isomer is observed. We believe

that the mer,cis-isomer of 7 is thermodynamically disfavored

because in such a structure a hydride ligand would be trans to

the oxo ligand. The oxo ligand in 7 is the strongest trans

influence ligand, and the preferred structure has the amido ligand

(the weakest trans influence ligand) trans to it. The isomerization

of 4 into 6 is a double silyl migration to O, and presumably

requires the migrating silyl group to approach the oxygen atom.

While 7a is an intermediate in the Mg/H₂reduction of 4a,

pure samples of 7a failed to react with Mg powder (same grade

as used for the reduction of 4) in THF or ether under 1 atm of

H₂. We propose that the presence of MgCl₂in solution is

necessary for the reduction of the oxo group by Mg⁰. In the

Mg/H₂reduction of 4, 1 equiv of MgCl₂is produced from the

initial reduction of the Re-Cl bonds.

The tetrahydrides 5 are operationally unsaturated complexes

in the sense that they can only achieve an 18e count by accepting

π-donation from the amido N. Such π-donation, however, does

not completely quench the electrophilic, unsaturated character

of the Re center. Indeed, 5a was found to readily form adducts

with a number of 2e ligands (Scheme 5). This may be viewed

as competition between the lone pair of N and an external ligand

L for the empty orbital at Re. With all three compounds 5a-c,

The identity of 10a was confirmed by a single-crystal X-ray

study (vide infra). The quality of the X-ray data was sufficient

to locate the hydrides and determine that 10a is a classical

tetrahydride. In the ³¹P{¹H} NMR, a doublet is observed for

the PNP phosphorus atoms (intensity of 2) and a triplet for the

P of PMe₃(intensity of 1). The overall apparent symmetry of

1

the molecule at 22 °C according to the H NMR is C_2V. The

hydride signal is extremely broad at 20 °C (not observed). At

-76 °C, the signal is decoalesced into three broadened

resonances at -1.38, -2.20, and -9.30 ppm in 1:2:1 ratio.

These NMR observations are thus consistent with the structure

found by X-ray diffraction in the solid state (vide infra) where

the two Re-H-(cis)H planes are approximately perpendicular.

Such symmetry is also supported by DFT calculations on the

model compound (vide infra).

The solution NMR data on 11a are also consistent with the

proposed formulation. A triplet (1P) and a doublet (2P) are

1

observed in the ³¹P NMR. The H NMR data are consistent

with the overall C_2Vsymmetry. The hydridic signal appears at

-11.24 ppm as a doublet of triplets (J_{HP(of PMe3)}) 75 Hz, J_HP(of

^PNP)) 11 Hz). The 11 Hz coupling to the P atoms of the PNP

ligand is typical for a hydride cis to a phosphine. The 75 Hz

coupling to P of PMe₃is rather unusual. It is well-known that

the magnitude of the two-bond coupling constant between H

and P (or P and P) attached to a metal center increases with the

H-M-P angle increasing from 90° to 180°. Examples of

complexes where the H-M-P angle is substantially smaller

9

J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004 6365

A R T I C L E S

Ozerov et al.

Scheme 6

Table 1.^a

R

13

14

15

16

17

18

19

20

21

Cy

ⁱPr

13a

13b

13c

n

14b

n

15b

n

16a

16b

16c

17a

17b

NR

18a

18b

NR

19a

n

NR

20a

n

NR

n

21b

n

^tBu

a

NR ) cannot be prepared according to Scheme 6. n ) preparation

not attempted. Numbers, e.g., 13a, correspond to compounds that were

observed or isolated.

whereas complexes 5 provide an easily accessible empty orbital

and can react by an associative mechanism.

Reactivity of 5 with Acyclic Olefins. Our results on the

reactivity of ethylene with 5a have been communicated.³⁷Here

we report on extended studies including higher alkenes and 5b

and 5c as well. In the presence of 4 or more equivalents of

alkene, polyhydrides 5 are transformed into an equilibrium

mixture of five-coordinate hydrido carbynes 13-20 and their

six-coordinate olefin adducts 13*-20* (Schemes 6 and 7, Table

Scheme 7

than 90° are quite rare. Rothwell reported^40,41a series of Ta

complexes 12 in which the H-Ta-P angle of ca. 65° results

in a J_HPof 75 Hz, exactly the value we observe for 11a. In the

case of Rothwell’s Ta d⁰hydrides the distortion away from the

octahedral geometry was ascribed to the strengthening of the

Ta-H bonds and of the π-bonding between Ta and the π-donor

trans to PR₃. We believe that in 11a, hydrides are similarly

distorted away from N and toward PMe₃and the diminished

H-Re-P angle gives rise to an unusually high J_HP. The

distortion of the hydrides in 11a probably has the same origin

as in 12, serving to enhance the Re-H bonds and the Re-N

bonding. Such distortion is also observed in 5. In fact, 11a can

be viewed as a product of replacement of the two hydrides in

the PNP plane in 5a with a PMe₃ligand.

1). Qualitatively, the rate of the reaction is primarily governed

by steric factors. For compounds 5a,b, the reaction with even

a stoichiometric amount of ethylene is complete in <5 min at

ambient temperature; reactions with monosubstituted alkenes

are also fast. Reactions with a 20-fold excess of 3-hexene

proceed to completion in ca. 2 h at ambient temperature, while

neither 5a nor 5b react with even neat C₂Me₄at 110 °C. 5c

reacts much more slowly than 5a,b, and the reaction proceeds

only with ethylene and monosubstituted alkenes. For example,

the reaction of 5c with ethylene (1 atm C₂H₄) requires ca. 24 h

to go to completion at ambient temperature, and the reaction of

5c with 4-chlorostyrene requires ca. 10 h at 110 °C. 5c does

not react with larger alkenes even at 160 °C.

The solution NMR data for compounds 13-20 and 13*-

20* are consistent with their formulation as rhenium carbynes.

The carbyne carbons resonate in the 260-280 ppm region of

the ¹³C NMR spectrum and show J_C-Pcoupling of 11-12 Hz.

The carbon chemical shift of the six-coordinate adducts is

typically 10-20 ppm upfield of the five-coordinate hydrido

carbynes. For compounds 13-20, the hydride resonates (as a

triplet) around -8 to -10 ppm. The hydride peak in 13b* was

found to be shifted downfield (2.21 ppm) compared to 13b; a

similar chemical shift was observed for the hydride in 16b*

(2.00 ppm). In five-coordinate carbynes, there is no ligand

directly trans to the hydride, while in the six-coordinate adducts

the hydride is trans to the coordinated alkene. It is well-

known^43-47that hydrides trans to an empty site generally display

upfield chemical shifts. The ³¹P NMR shows a singlet for 13-

20 (except 19), and selective decoupling of the alkyl hydrogens

converts such singlets into doublets, showing hydride coupling.

Notably, the conversion of 5a to 11a is a replacement of 2 H

(lost as H₂) by a PMe₃ligand achieved under very mild

conditions. For most polyhydrides, replacement of 2H by PR₃

does not proceed readily and requires elevated temperature⁴²

and/or a sacrificial olefin to scavenge 1 equiv of H₂by hydro-

genation. The kinetic difficulty lies in the fact that polyhydrides

of a general formula (R₃P)_xMH_yare invariably saturated (18e)

compounds and must react by a dissociative mechanism,

(40) Clark, J. R.; Pulvirenti, A. L.; Fanwick, P. E.; Sigalas, M.; Eisenstein, O.;

Rothwell, I. P. Inorg. Chem. 1997, 36, 3623.

(43) Esteruelas, M. A. W.; H. J. Organomet. Chem. 1986, 303, 221.

(44) Werner, H.; Wolf, J.; Hohn, A. J. J. Organomet. Chem. 1985, 287, 395.

(45) Intille, G. M. Inorg. Chem. 1979, 11, 695.

(41) Parkin, B. C.; Clark, J. C.; Visciglio, V. M.; Fanwick, P. E.; Rothwell, I.

P. Organometallics 1995, 14, 3002.

(46) James, B. R.; Preece, M.; Robinson, S. D. Inorg. Chim. Acta 1979, 34,

L219.

(42) Smith, K. J.; Ondracek, A. L.; Gruhn, N. E.; Lichtenberger, D. L.; Fanwick,

P. E.; Walton, R. A. Inorg. Chim. Acta 2000, 300-302, 23.

(47) Masters, C.; Shaw, B. L. J. Chem. Soc. A 1971, 3679.

9

6366 J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004

Transformation of Acyclic Alkenes to Hydrido Carbynes

A R T I C L E S

The position of the equilibrium between a five-coordinate

carbyne and its olefin adduct (Scheme 7) also depends primarily

on steric factors. For derivatives of the PNP^tBuligand, no detect-

able six-coordinate adduct was observed even in the case of

ethylene. As we reported previously, for 13a the equilibrium

lies strongly on the side of the olefin adduct 13a*. The PNP^iPr

analogue 13b behaves similarly. The six-coordinate adducts of

smaller monosubstituted olefins are observable in solution, but

lose coordinated olefin readily upon exposure to vacuum at

ambient temperature. The larger neo-hexene binds very weakly,

causing only a broadening of the NMR resonances of the five-

coordinate carbyne when present in large excess (ambient tem-

perature). Alkene adducts were not observed for any di- or tri-

substituted alkenes. Upon dissolution in C₆D₆(ca. 22 °C), an

equilibrium mixture of 16b, 16b*, and 4-chlorostyrene is formed

in the ratio of 1:10:1 (¹H and ³¹P NMR integration). K_eqof

10 corresponds to ∆G of binding of 4-chlorostyrene of ca.

1.4 kcal/mol.

With straight chain alkenes, either terminal or internal, there

is only one possible carbyne to be formed, where the metal is

attached to the terminus of the hydrocarbon chain. In branched

alkenes, there may be inequivalent chain termini, and thus more

than one carbyne can be formed. Indeed, in the case of the

reaction of 5a with 2-methylbutene, two carbyne products (19a,

20a) were identified by NMR. 19a must have C₁symmetry as

the â-carbon has four different substituents. This lack of

symmetry is consistent with our observation of an AB pattern

(50.1, 50.3 ppm) in the ³¹P NMR spectrum for one of the

carbynes formed from 5 and 2-methylbutene and of a dissym-

metric picture in its ¹³C NMR. The J_P-Pcoupling constant in

19a is 205 Hz, consistent with the approximately trans geometry

of the two phosphine arms. The ca. two times smaller value of

J_P-Pfor the alkene adducts 16* and 18* probably reflects the

compression of the P-Re-P angle as a consequence of

accommodation of an alkene ligand. For instance, the P-Re-P

angle in the solid-state structure of 17a is ca. 167° (vide infra),

while for the adduct of even the smallest alkene, 13*a, it is

only 157.5°. It is reasonable to expect that alkenes larger than

ethylene such as 1-hexene in 18* or 4-chlorostyrene in 16* will

cause an even greater compression of the P-Re-P angle.

The five-coordinate carbynes 13-20 are all either viscous

oils or extremely lipophilic solids, which prevented isolation

of most of them as analytically pure solids. We were able,

however, to obtain analytically pure, crystalline samples of 17a,

including a crystal suitable for an X-ray diffraction study (vide

infra). Compounds 13-20 are also quite thermally stable. No

decomposition could be detected after thermolysis (110 °C, 24

h) in a protio-alkane or arene solvent for 13a-c, 16a-c, and

17a-b (others were not tested). No noticeable H/D exchange

with C₆D₆occurs in solutions of 13-20 at ambient temperature

on the time scale of 24 h. Prolonged heating of C₆D₆solutions

does result in some H/D exchange as evidenced by the decrease

in intensity of the hydride signal and the PNP resonances as

well as the change in multiplicity for the latter.

Alkene binding competes with π-donation from the amido

lone pair for the empty orbital at Re. Similarly to the tetrahy-

drides 5, five-coordinate carbynes 13-20 are also operationally

unsaturated complexes. The color is very indicative: operation-

ally unsaturated complexes tend to be deeply colored (carbynes,

red; tetrahydrides 5, purple-red) while classically saturated

adducts are colorless (e.g., 10a, 13*-20*). Thus, dissolution

of the colorless 16b* leads to a reddish solution as a result of

an equilibrium concentration of 16b. The temperature-dependent

nature of the equilibrium is evident from the gradual disap-

pearance of color from solutions of 16b* upon cooling and its

reversible reappearance upon warmup.

The fact that hydrido carbynes were formed even from inter-

nal alkenes such as 3-hexene shows that, in some intermediate,

the metal has the ability to migrate along the hydrocarbon chain

to the terminus in order to form a carbyne product. The reaction

of 5b with excess (20 equiv) 3-hexene after 2 h at ambient

temperature produces in situ a mixture of five-coordinate 18b

and two other products, each showing an AB pattern in the ³¹

P

Reactivity of 5b with Allylamine. With allylamine as a

substrate we found that the reaction eventually gives the carbyne

product, as it does with hydrocarbyl alkenes. After 2 h at

ambient temperature, the reaction of 4 equiv of allylamine with

5b gave an essentially colorless mixture of two diastereomers

of 21*b. Application of vacuum to this solution caused

reddening; however, when the volatiles were completely

stripped, the residual solid was of an off-white color. This solid

was poorly soluble in aliphatic hydrocarbons yet dissolved

readily in C₆D₆, giving a red solution. Stripping of the volatiles

generated a colorless solid again. The solution of this colorless

NMR spectrum in the 41-44 ppm region. Removing the vola-

tiles in vacuo causes the AB ³¹P signals to disappear, and the

residue contains only 18b (by NMR). Addition of excess

3-hexene to pure 18b does not lead to changes observable by

NMR. On the other hand, addition of 1-hexene to pure 18b

leads to the complete disappearance of the NMR signals of 18b

and the appearance of the two AB patterns observed in situ in

the reaction of 5b with 3-hexene. Removal of volatiles in vacuo

regenerates pure 18b. We conclude that these two pairs of AB

patterns belong to the two possible diastereomers of a 1-hexene

adduct of 18b (18b*, Scheme 7, R′ ) n-Bu). Presumably, the

alkene C-C bond is approximately aligned with the P-P vector

in the six-coordinate adducts 13*-20*, as was found (see

below) in the solid-state structure of 13a*. This can lead to two

diastereomers for monosubstituted alkenes: the substituent on

the olefin is either syn or anti to the carbyne. In either of these

isomers the phosphines are inequivalent, giving rise to AB

patterns. Observation of only one such isomer for 16b* (a pair

of AB doublets, 36.5 and 41.7 ppm, J_P-P) 99 Hz) indicates a

greater energy difference between the two diastereomers, likely

because of the increased steric demand of an aryl substituent

compared to an n-alkyl. The 1-hexene adducts 18a* were also

observed in the reaction of 5a with 3-hexene.

1

solid displayed somewhat broad signals in H, ¹³C, and ³¹P

NMR, consistent with its formulation as 21b. No free or bound

allylamine was detected in the solution of this solid, and

elemental analysis confirmed the formulation of this solid as

21b. We propose that an equilibrium exists between a mono-

meric 21b (red) and its dimer (or oligomer) in which the

dangling amino group of the carbyne moiety serves as a ligand

to another molecule of 21b forming a (colorless) adduct. In

solution, the equilibrium lies on the side of the monomer, but

in the solid state the di- or oligomerization prevails, judging by

the (absence of) color and the decreased solubility. Compound

21b is indefinitely stable in solution at ambient temperature,

but decomposes slowly to unidentified products when heated

9

J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004 6367

A R T I C L E S

Ozerov et al.

produced (NMR evidence). Notably, neither EtF (¹⁹F NMR) nor

EtOPh (¹H NMR) were observed in the reactions of vinyl

fluoride or vinyl phenoxide with 5b. Unfortunately, the extreme

lipophilicity of 22b did not allow its isolation in a pure solid

form, and it was characterized by solution NMR. At 22 °C, a

³¹P{¹H} NMR singlet is observed, which changes to a broad

triplet upon selective decoupling of alkyl hydrogens. In the ¹H

NMR, a broad peak integrating to 2H at -9.89 ppm is observed,

consistent with the presence of two H bound to Re. The other

¹H and ¹³C signals at 22 °C are consistent with an average C_s

symmetry. The signals corresponding to the EtO group are

observed in the ¹H NMR, with the CH₂group being considerably

shifted downfield (4.31 ppm) from its position in Et₂O or

EtOCHdCH₂. This shift may be indicative of substantial

involvement of the oxygen in stabilizing the empty orbital of

the carbene. In the ¹³C NMR, the carbene R-C is observed as a

triplet (J_PC) 8 Hz) at 273.1 ppm.

Since for a static 22b, equivalence of both the P nuclei and

the two Re-H does not seem possible, we undertook a low-

temperature NMR study. At -60 °C, the two hydride signals

are decoalesced into two well-separated broad peaks. At the

same time, only one ³¹P resonance is observed throughout the

range of -60 to 22 °C, as well as only one set of EtO resonances

in the ¹H NMR. There are several possibilities for the structure

of 22b. The carbene ligand may be either “in between” the two

hydrides or “on the side” of them (C and S). For the structure

S, it is also possible to have either a classical dihydride or a

dihydrogen complex. It is unlikely that any minimum corre-

sponds to the rotamer in which the plane of the carbene is in

the plane of the PNP ligand because in that case we would

expect to observe inequivalent ³¹P nuclei. Both the (static)

structures C and S should give rise to two inequivalent hydride

signals. The rotation about the Re-C bond can serve to time-

average the hydride signals at 22 °C in C. For S, the two

hydrides may be exchanging by a rotation of an intermediate

dihydrogen isomer. We currently favor structure S based on

the DFT study of a model system (vide infra).

Scheme 8

to 110 °C. The lower thermal stability of 21b compared to 13-

20 is probably caused by some degradation of 21b by the amino

functionality at elevated temperatures.

Reactivity of 5 with Heteroatom-Substituted Alkenes.

Following the long standing interest in the transformations of

heteroatom-substituted alkenes on strongly reducing transition-

metal centers,^4,8we also investigated the reactivity of 5 with

such substrates. While the ambient temperature reactivity of

allylamine parallels that of the simple hydrocarbyl alkenes, this

was not found to be the case with vinyl ethers and vinyl fluoride.

Addition of vinyl phenyl ether (4 or more equivalents) to

solutions of 5b in an arene or alkane solvent results, in the time

of mixing, in the formation of a colorless solution. The ³¹P NMR

analysis of the resultant mixture revealed the presence of two

sets of AB patterns. If this solution was allowed to stand for

several hours, a significant amount of decomposition to

unknown products was observed. The chemical shift and the

J_P-Pcoupling constant for the observed pair of AB patterns were

consistent with the formation of two diastereomers of a six-

coordinate adduct of an alkene with a hydrido carbyne (13b*OPh,

Scheme 8). Removal of the coordinated alkene in vacuo

proceeded with difficulty to produce the red five-coordinate

carbyne. The NMR resonances for this carbyne were consistent

with it being 13b. Thus, in the course of the reaction the C-O

bond was cleaved. Addition of vinyl phenyl ether to an authentic

sample of 13b reproduced the ³¹P NMR observation of two AB

patterns. The difficulty in removing the alkene in vacuo is then

simply explained by the low volatility of the alkene being

removed: vinyl phenyl ether.

The reaction of vinyl fluoride with 5b proceeded in a similar

fashion to give a solution containing a major product (ca. 85%)

displaying an AB pattern in ³¹P NMR. Compared to the reaction

with vinyl phenyl ether, the reaction with vinyl fluoride resulted

in a slightly greater amount of decomposition products. Heating

this product mixture in vacuo resulted in the disappearance of

the AB pattern and the formation of 13b as the major component

Thermolysis of 22b (30 min, 80 °C) in the presence of ethyl

vinyl ether produces five-coordinate 13b, with some concomitant

decomposition. The NMR signals of 13b in the presence of an

excess of ethyl vinyl ethers are only slightly broadened,

indicating that ethyl vinyl ether binds to 13b more weakly than

vinyl phenoxide, vinyl fluoride, or other monosubstituted alkenes

(vide supra). The conversion of 22b to 13b must proceed with

loss of EtOH. Similarly, PhOH and HF are the expected

coproducts in the reactions of 5b with vinyl phenoxide and vinyl

fluoride. The cleavage of Si-N bonds of the PNP ligand by

ROH and HF could account for the partial decomposition

observed in these reactions. To test this hypothesis, we

conducted the reactions of Scheme 8 in the presence of a 20-

fold excess of Me₃SiNMe₂as a source of sacrificial Si-N bonds.

In the case of vinyl ethers, solutions of 13b*OPh and 13b thus

formed were >98% pure and persisted for at least 48 h without

any sign of decomposition. The reaction of 5b with vinyl

1

(identified by ³¹P, ¹³C, H NMR in situ), the product of C-F

bond cleavage. This AB pattern corresponds to the major

diastereomer of 13b*F, the vinyl fluoride adduct of 13b.

Because of the low concentration of the minor isomer of 13b*F,

we were unable to identify its ³¹P NMR resonances. However,

the ¹⁹F NMR signals of both isomers were identified, and the

assignment was confirmed by observation of only these two

¹⁹F NMR peaks when vinyl fluoride was introduced into a

solution of an authentic sample of 13b.

The reaction of 5b with excess vinyl ethyl ether at ambient

temperature proceeded rapidly to give, as the intermediate

product, orange-red 22b (Scheme 8). Concomitantly, Et₂O was

9

6368 J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004

Transformation of Acyclic Alkenes to Hydrido Carbynes

A R T I C L E S

Table 2. Selected Bond Distances and Angles in the Solid-State

Structures of 10a and 17a (Two Independent Molecules)

10a

17a, 1st molecule

17a, 2nd molecule

Bond Distances, Å

Re-N

2.2503(13)

2.4187(4)

2.4089(4)

2.3478(4)

1.55(3)

1.59(3)

1.60(3)

1.63(2)

-

2.112(3)

2.3868(8)

2.3847(8)

-

2.112(3)

2.3896(9)

2.3852(9)

-

Re-P1

Re-P2

Re-P3

Re-H1

Re-H2

Re-H3

Re-H4

Re-C31

N-Si1

N-Si2

H1-H2

H3-H4

1.60

-

1.761(3)

1.737(3)

1.722(3)

-

1.754(3)

1.731(3)

1.725(3)

-

1.6964(13)

1.6915(14)

1.88

Figure 1. ORTEP drawing (PNP^Cy)Re(H)₄(PMe₃) (10a) showing selected

atom labeling. All hydrogen atoms not attached to Re and all methylene

carbons of the Cy rings are omitted for clarity.

1.77

-

Bond Angles, deg

fluoride in the presence of a large excess of Me₃SiNMe₂gave

a substantially purer 13b*F, although some side products were

still observed. The observation of Me₃SiF by ¹⁹F NMR in the

latter reaction was crucial, supporting the proposed generation

of HF.

P1-Re-P2

N-Re-P3

143.242(13)

153.36(4)

-

167.34(3)

-

139.36(14)

127.1

167.02(3)

-

139.66(14)

126.6

N-Re-C31

H1-Re-N

H1-Re-C31

Re-C31-C32

-

92.3

169.4(3)

93.7

171.0(3)

Removal of volatiles in vacuo from solutions of pure 22b

results in partial conversion of 22b to 13b but also to 23b. We

propose the structure of 23b based on the solution NMR

evidence. The ³¹P NMR singlet of 23b resonates at ca. 61 ppm,

in the region where signals of 13b-21b are found. Selective

decoupling of aliphatic hydrogens results in a doublet in the

³¹P NMR spectrum, consistent with the presence of a single

hydride on Re. Because 23b was only observed as a minor

component in the presence of other compounds, we were unable

1

to obtain full H and ¹³C NMR data. However, the chemical

shift and the value of J_HP(J_CP) for the ¹H hydride signal of 23b

and for the signal of the R-C were very similar to those of 13-

21. In addition, the signals arising from the two CH₂groups

Figure 2. ORTEP drawing (PNP^Cy)ReH(≡CCH₂CMe₃) (17a) showing

selected atom labeling. All hydrogen atoms not attached to Re and all

methylene carbons of the Cy rings are omitted for clarity.

1

connected to O in 23b were also identified in the H and ¹³C

NMR spectra.

23b was also observed when 5b was allowed to react with 2

equiv of ethyl vinyl ether at 22 °C; after 2 h, the mixture

consisted of (³¹P NMR) 5b, 13b, 22b, and 23b in a 6:1:75:18

ratio. However, upon standing at 22 °C, the ratio changed to

11:19:69:1. Apparently under the reaction conditions, 23b can

be converted to 13b, although at the present time the mechanism

of this transformation is unclear.

of 5a where the Re center can achieve the 18-electron config-

uration only by accepting π-donation from N. The Re-N bond

in 5a (ca. 2.06 Å) is much shorter than in 10a (2.25 Å). At the

same time, the average N-Si bond in 5a (ca. 1.745 Å) is longer

than in 10a (ca. 1.694 Å) while the sum of angles about N is

ca. 360° in both cases. The π-basic N of the PNP ligand is

connected to three competing π-acids (Re and 2 Si). The longer

Re-N bond in 10a compared to 5a is not simply a consequence

of the higher coordination number of 10a. In fact, the Re-P

distances in 10a are slightly shorter than those in 5a.

Solid-State Structure of 17a (Figure 2, Table 2). The

presence of a carbyne ligand is confirmed by the short Re-C

bond comparable to other Re carbynes and by the near linear

geometry of the Re-C31-C32 fragment. The coordination

environment about the Re center in 17a is most appropriately

described as Y-shaped five-coordinate. The two phosphine

donors are approximately trans to each other, while H1, N, and

C31 form a Y-shaped geometry about the Re center in the

perpendicular plane. The Y-shaped geometry is defined by an

acute angle of 90° or less (here: H1-Re-C31) and two other

angles of >120° (here: N-Re-C31, H1-Re-N). The overall

geometry about Re is closely reproduced by the DFT calculation

on the model compound 13^H, including the position of the

hydride ligand. The Y-shaped geometry can be understood as

derived from a square pyramidal geometry with the π-donor

Solid-State Structure of 10a (Figure 1, Table 2). A crystal

of 10a suitable for an X-ray diffraction study was obtained by

treating a pentane solution of 5a with excess PMe₃and cooling

the resultant solution to -30 °C. The quality of the X-ray

diffraction data was sufficient to locate the hydride ligands, and

these data show 10a to be a classical tetrahydride. The Re center

in 10a is eight-coordinate, with the heavy atom donors forming

a strongly flattened tetrahedral arrangement. The donor atoms

are arranged in two approximately perpendicular planes: the

first contains P1, P2, H1, H2 and bisects the H3-Re-H4 angle;

the second contains P3, N, H3, H4 and bisects the H1-Re-

H2 angle. The dodecahedral structure avoids placement of the

strong trans influence hydrides trans to each other. The Re-P3

distance is somewhat shorter than the Re-P1 and Re-P2

distances, likely a consequence of the smaller cone angle of

PMe₃. The key feature of the structure of 10a is the long Re-N

bond. Complex 10a possesses an 18-electron count without the

need for the π-donation from the amido N; therefore, the Re-N

bond in 10a is a single bond. This is in contrast to the structure³⁷

9

J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004 6369

A R T I C L E S

Ozerov et al.

Scheme 9

(here: N of PNP) shifting to a position that maximizes

π-donation to the metal center.

We have previously reported³⁷the structure of the six-

coordinate 13a*, and now we report the structure of the five-

coordinate, operationally unsaturated hydrido carbyne 17a. We

believe that the structure of 17a is representative of the structures

of 13-21. The presence of the π-donation from N to Re in

17a, but not in 13a*, is evinced by the much shorter Re-N

bond in 17a compared to that in 13a* (ca. 2.28 Å) and the

reverse trend in N-Si distances. Either of the structural

comparisons between 5a and 10a or between 17a and 13a* is

a comparison between an operationally unsaturated compound

with a multiple, short Re-N bond and its saturated adduct with

a long Re-N (single) bond.

of 21b by liberated H₂is higher than for 16b because H₂has to

compete with the pendant amino group for coordination to Re.

This reaction scheme implies that hydrogenation of a six-

coordinate hydrido carbyne or even of a five-coordinate hydrido

carbyne is slow. Indeed, exposure of a solution of 13a* to 1

atm of H₂(excess H₂, C₆D₆) does not lead to any immediate

changes (by ¹H and ³¹P NMR). Upon heating of this sample, it

was found that 13a* is completely converted to 5a only after

>10 h at 90 °C. When the five-coordinate 13a was exposed to

excess H₂, the complete hydrogenation to 5a proceeded faster

(10 h, 22 °C) than in the case of 13a*, but was still much slower

than the time scale of formation of hydrido carbynes. Excess

alkene suppresses the hydrogenation of the hydrido carbyne

complexes because H₂has to compete with alkene for coordina-

tion to Re.

The conversion of 5 to 13-21 is clearly a multistep

transformation.^48-54The proposed mechanism based on our

observations and on literature precedents is outlined in Scheme

10. Initial coordination of alkene to 5 is followed by insertion

into one of the Re-H bonds and reductive elimination of the

corresponding alkane from 24. Coordination of another alkene

molecule may occur before or after alkane elimination from

24, furnishing the alkene dihydride 25. 25 can isomerize to the

carbene dihydride 28, which, with concomitant loss of H₂,

evolves into the hydrido carbyne product (13-21).

Discussion

Transformations of Hydrocarbyl-Substituted Alkenes. The

π-donating ability of the amido N contributes to the ability of

the PNP^Rligand to sustain operationally unsaturated complexes.

This is demonstrated by the facile and reversible binding of

olefins by 13-20,³⁷of H₂by 5 in the reactions of Scheme 5,

and by the facile replacement of H₂in 5 by PMe₃. Therefore,

it is likely that the activation energy of binding of an olefin to

5 is quite small at ambient temperature. Hydrido carbyne

complexes 13-21 contain the (PNP^R)Re fragment and the

elements of the corresponding alkene. It is clear then that,

formally, the four hydrides of 5 are removed from, and the

elements of the alkene added to, Re to give the hydrido carbyne

product. The obvious proposal is that the four hydrides are spent

in the hydrogenation of 2 equiv of alkene to the corresponding

alkane. Thus, 3 equiv of alkene would be needed to produce a

five-coordinate hydrido carbyne (13-20) from 1 equiv of 5,

and the fourth equivalent is necessary to produce an alkene

adduct (13*-21*).

We believe that the initial insertion is quite fast as demon-

strated by an experiment with a relatively bulky alkene. When

a heptane solution of 5b was treated with 3 equiv of cis-stilbene

(mp ) -5 °C), a copious amount of trans-stilbene (mp ) 122

°C) precipitated within ca. 2 min while still only and all of 5b

was observed in solution by ³¹P NMR. The isomerization most

likely proceeds by insertion/â-H-elimination. Reaction of trans-

stilbene with 5b proceeds only very slowly; at 100 °C, more

than 24 h is required to consume 5b by 3 equiv of trans-stilbene.

Stilbene is not an olefin that can be isomerized to a hydrido

carbyne, and thus this reaction yields a mixture of products

(including bibenzyl) that apparently include products of activa-

tion of the solvent. The sluggishness of the hydrogenation of

stilbene by 5b should probably be attributed to its relative steric

bulk (cf. Scheme 6). That may be an indication that reductive

elimination from 24 is assisted by the incoming olefin.

However, the reality is less straightforward. If an excess (20

equiv) of BuCHdCH₂is allowed to react with 5b in a closed

t

NMR tube in C₆D₆, then 17b is produced quantitatively, but

t

only 1 equiv (by ¹H NMR integration) of BuCH₂CH₃ac-

companies its formation. In another experiment, 3 equiv of

p-chlorostyrene was reacted with 1 equiv of 5b in a closed tube

in protio-cyclohexane. After 2 h at ambient temperature, ³¹P

NMR indicated that the reaction mixture consisted of >90%

16b*, the remainder being the unreacted 5b and the five-

coordinate 16. After 24 h at ambient temperature, the composi-

tion of this sample changed to contain ca. 50% of 16, and after

heating this mixture at 100 °C for 2 h, 16 became the dominant

(>90%) component. Similar results were obtained in the

reactions of allylamine with 5b. With 2 equiv allylamine (closed

tube), after 30 min at ambient temperature 21b was the major

(∼90%) component and remained the major component after

48 h at 22 °C. If excess allylamine was allowed to react with

5b, then 21b* was the major product observed. These experi-

ments indicate that under kinetic control only 2 equiv of alkene

are needed to produce a five-coordinate hydrido carbyne, and

3 equiv to produce its six-coordinate alkene adduct. We propose

that under kinetic control, two of the original four hydrides of

5 are released as free H₂. This H₂can, over longer reaction

times and/or higher temperature, hydrogenate another equivalent

of alkene, but only after most of Re is converted to a hydrido

carbyne (Scheme 9). Presumably, the barrier to hydrogenation

(48) Shih, K.-Y.; Totland, K.; Seidel, S. W.; Schrock, R. R. J. Am. Chem. Soc.

1994, 116, 12103.

(49) Schrock, R. R.; Seidel, S. W.; Mosch-Zanetti, N. C.; Shih, K.-Y.;

O’Donoghue, M. B.; Davis, W. M.; Reiff, W. M. J. Am. Chem. Soc. 1997,

119, 11876.

(50) Veige, A. S.; Wolczanski, P. T.; Lobkovsky, E. B. Angew. Chem., Int. Ed.

2001, 40, 3629.

(51) Freundlich, J. S.; Schrock, R. R.; Davis, W. M. J. Am. Chem. Soc. 1996,

118, 3643.

(52) Hughes, R. P.; Maddock, S. M.; Rheingold, A. L.; Guzei, I. A. Polyhedron

1998, 17, 1037.

(53) Miller, G. A.; Cooper, N. J. J. Am. Chem. Soc. 1985, 107, 709.

(54) Giannini, L.; Guillemot, G.; Solari, E.; Floriani, C.; Re, N.; Chiesi-Villa,

A.; Rizzoli, C. J. Am. Chem. Soc. 1999, 121, 2797.

9

6370 J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004

Transformation of Acyclic Alkenes to Hydrido Carbynes

A R T I C L E S

Scheme 10

from an internal position. Two possibilities can be envisioned

(Scheme 12). If 25 and 28 (Scheme 10) can indeed interconvert

Scheme 12

under the reaction conditions, then Re can traverse the skeleton

without releasing free alkene. An alternative option is that 5

mediates fast isomerization of alkenes, thus producing an

equilibrium concentration of free terminal alkene isomer, which

is then transformed into the carbyne. The latter option is

supported by the observation that the reaction of 5 with 3-hexene

results in the formation of a mixture of 18 and 18*, i.e., some

1-hexene is generated beyond what may be needed to produce

the carbyne. In addition, the potential of 5 for isomerizing

alkenes is supported by the cis- to trans- isomerization of stilbene

(vide supra). However, we cannot rule out that both pathways

are operative.

Transformations of Heteroatom-Substituted Alkenes. Re-

actions of H₂CdCH-X (X ) F, OPh, OEt) also lead to the

hydrido carbyne products, thus testifying to the high preference

of the (PNP^R)Re fragment for the formation of carbyne

compounds. However, the difference from the reactions of

hydrocarbyl-substituted olefins is that the C-X bond is cleaved

in the process (Scheme 13), resulting only or primarily in

None of the species 26-28 were observed, but in certain cases

we were able to observe analogues of 25 (Scheme 11). Allowing

Scheme 11

5b to react with 2 equiv of p-chlorostyrene in cyclohexane-h₁₂

produces a mixture of 16b and 5b after 5 h. However, before

this conversion is complete, a substantial amount of the species

29 was observed. 29 is characterized by an AB pattern in the

³¹P NMR spectrum (J_PP) 92 Hz), and selective decoupling of

the aliphatic hydrogens results in the splitting of each line in

this AB system into a triplet. This is indicative of the presence

of two hydrides on the Re center. No upfield ¹³C resonance

assignable to a carbene (such as in 28) carbon was detected by

¹³C NMR under conditions (concentration, time) where the

carbene signal of 22b or carbyne signals of 13-20 were

unambiguously observed. These data are consistent with the

proposed identity of 29. Binding of a monosubstituted olefin

will render the two ³¹P nuclei in the molecule inequivalent

regardless of the orientation of the olefin. We tentatively favor

the structure of 29 where the CdC vector of the olefin is

approximately aligned with the P-M-P vector based on the

DFT studies of the model compound (vide infra). Because 29

is unstable in solution and transforms into 16, we were unable

to obtain it in a pure form in solution or in the solid state. Similar

observations were made in the reaction of 5b with allylamine.

In addition to peaks of 21 and 21b*, the ³¹P AB system (J )

178 Hz) of compound 30 was observed at shorter (<60 min)

reaction times, which then transformed into 21b. Compounds

29 and 30 were also observed (<1 h reaction time) in the

reactions of 5b with 3 equiv of p-chlorostyrene and allylamine

(vide supra), respectively.

Scheme 13

ethylidyne products.

In addition, we observed that the reactivity of H₂CdCH-

OEt is different from the reactivity of H₂CdCH-OPh and

H₂CdCH-F, as only in the case of ethyl vinyl ether did we

see (vide supra) (a) the product of the hydrogenation of the

double bond, (b) the observable and isolable alkoxycarbene

intermediate 22, and (c) the alkoxymethylcarbyne 23b as a

minor kinetic product. It seems reasonable to propose (Scheme

14) that initially coordination of H₂CdCH-X to 5 takes place,

followed by insertion into a Re-H bond. Reductive elimination

at this stage with formation of a C-H bond accounts for the

formation of Et₂O (X ) OEt). When X ) F or OPh, reductive

elimination from the â-X-substituted alkyl trihydride does not

occur as Et-X is not observed for X ) F, OPh. â-X-elimination

followed by HX reductive elimination may be suggested as a

competitive process. This would produce an ethylene complex

which is then rearranged to hydrido ethylidyne as discussed

Isomerization of Internal Alkenes. The hydrido carbynes

18-20 are produced from internal alkenes. In the carbyne

complexes, Re is attached to a terminus of the hydrocarbon

skeleton, and thus the question arises of how Re migrates there

9

J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004 6371

A R T I C L E S

Ozerov et al.

Figure 3.

Scheme 14

in the 1,2 and 2,1 fashion. We propose here that the production

of both 13b and 23b from 5b and ethyl vinyl ether is a result

of two competitive modes of insertion of ethyl vinyl ether into

the Re-H bond of the (PNP)ReH₂fragment. Insertion followed

by R-H migration results in two different carbenes 22 and 31

(unobserved), which then produce 13 or 23 by loss of HX or

H₂, respectively. Interestingly, all the transformations and

intermediates shown in Scheme 14 are valid for X ) H as well

(cf. Scheme 10).

We have previously^2,3,5described the chemistry of the

cleavage of the C-X bond in vinyl-X compounds by (R₃P)₂-

Os(H)₃Cl. The major final Os products for X ) OPh or F

(although not for X ) OEt) were ethylidyne compounds (R₃P)₂-

OsHClX(≡CMe), but in the case of X ) F formation of ethylene

from C₂H₃F was also observed. Intermediate carbene complexes

(R₃P)₂OsHCl(dCMeX) were observed for X ) OPh or OEt.

DFT Computational Evaluation of Potential Intermedi-

ates. Although we have not surveyed the entire PES for the

carbyne synthesis reaction (in particular, we have not sought

transition states), we have evaluated the free energies of several

obvious potential intermediates. We used a model (H₂PCH₂-

above. Since â-X-elimination implies migration of the X^-anion,

OPh and F should be more prone to migrate than OEt. In the

language of organic chemistry, OEt is a worse leaving (i.e.,

anion-forming) group than OPh or F.

It was previously shown¹³that vinyl ethyl ether can reversibly

insert into the Ru-H bond of the (ⁱPr₃P)₂RuHCl fragment both

9

6372 J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004

Transformation of Acyclic Alkenes to Hydrido Carbynes

A R T I C L E S

Scheme 15

attributed to the strengthening of the metal-oxo π-bonding. This

is probably reinforced by the unfavorable nature of a 180°

H-Re-H angle because of the high trans influence of the

hydrides and by the chelate constraint that “pulls back” the

phosphine arms. The Re-N(amido) distance of 2.345 Å is the

longest among all molecules reported here. The oxo ligand

should be regarded as a stronger π-donor than the disilylamido

moiety of PNP. The 18e count at Re can be achieved by double-

face π-donation from the oxo ligand, without need for the p-lone

pair on N. The Re-N bond is thus a single bond. It is longer

compared to other single Re-N bonds reported here, reflecting

the high trans influence of the oxo ligand. The Re-O bond

distance is very similar to that calculated³⁹for 4^H.

Scheme 16

Scheme 17

Compound 10^H. The optimized structure of 10^Hreproduces

the idealized mirror symmetry of the structure of 10a obtained

in the X-ray diffraction and solution VT NMR studies (1:2:1

hydride signals). However, the X-ray diffraction-derived struc-

ture of 10a possesses four classical hydride ligands, whereas

the DFT-optimized structure of 10^Hpossesses one dihydrogen

ligand (distance H3-H4 is 0.924 Å) and two hydrides (the

structure with four hydride ligands is not a minimum on the

potential energy surface). While it would be tempting to cite

the unreliability of the X-ray diffraction for determination of

exact hydride positions, we feel such a simplistic rationale alone

is not sufficient. Besides the H-H distances, the DFT-optimized

structure of 10^Halso possesses P1-Re-P2 and P3-Re-N

angles that are larger than those in the X-ray diffraction structure

of 10a. It is consistent that to accommodate the increase in H-H

distances (and thus in H-Re-H angles), the P1-Re-P2 and

P3-Re-N angles should decrease. The origin of this discrep-

ancy is not clear, but it is possible that it can be traced to use

of the model PNP^Hligand (and PH₃instead of PMe₃) in our

DFT calculations. Compared to 10a, 10^Hhas seven out of nine

alkyl substituents on the three phosphine donors replaced by

hydrogens. This decreases the basicity of Re in the model system

and therefore its ability to reduce η²-H₂to two hydrides. In

addition, the steric bulk of the experimental ligand PNP^Rmay

favor the arrangement of heavy atom donors about Re that is

closer to tetrahedral (10a). The rehybridization of the Re

d-orbitals as a consequence of such distortion may lead to higher

stability of the dihydride structure (10a) vs the η²-H₂structure

(10^H). Whether a Re^IIIdihydrogen complex or a Re^Vtetrahy-

dride, 10 is an 18-electron complex without need for the p-lone

pair of the amido N. Accordingly, the Re-N bond in 10^His

calculated to be long at 2.210 Å, consistent with its single bond

character. The near-zero free-energy change of eq 1 is, as

discussed above:

Scheme 18

Scheme 19

Scheme 20

SiH₂)₂N ligand (PNP^H) in our DFT (B3PW91) studies. Such

model compounds will be denoted by the superscript “^H”. DFT

is a vital supplement to experimental studies not only because

it allows the strengthening of conclusions derived from experi-

ment, but also because it can provide valuable information with

respect to energetics of reactions.⁵⁵It is also often indispensable

as a source of information about energies and geometries of

compounds not observed experimentally or some features of

observed compounds evading experimental confirmation (e.g.,

hydride ligand positions). With that in mind, the two major

objectives of our DFT study were to investigate the structures

of the (PNP)Re complexes and to determine the relative energies

of key compounds (and therefore the energetics of pertinent

transformations). The structures of all the DFT-optimized

geometries are pictured in Figure 3 and with energies given in

Schemes 15-20 (∆G°₂₉₈in kcal/mol).

+1.2

[PNP]Re(H)₄+ PH_{3 kcal/mol}8 [PNP]Re(H)₂(H₂)(PH₃) (1)

consistent with considerable stabilization of the “operationally

unsaturated” [PNP]Re(H)₄by N f Re π-donation. Indeed, the

calculated Re-N distance increases by 0.15 Å from reactant to

product.

Compound 13^H. We chose 13^Has a model for the five-

coordinate hydrido carbynes 13-21. The DFT-optimized ge-

ometry of 13^Hreproduces the key parameters of the experi-

mentally determined structure of 17a very closely. This

agreement reinforces our confidence in the location of the

hydride ligand in the X-ray diffraction study and in the overall

Y-shaped geometry about Re.

Compound 7^H. The optimized structure of compound 7^H

(trans,trans-isomer) can be described as a strongly distorted

octahedron. The phosphine donors of the PNP ligand and the

hydrides strongly bend away from the oxo ligand. Such bending

away of ligands cis to an oxo (or nitrido) ligand has been

(55) ReactiVe Intermediate Chemistry; Moss, R. A.; Platz, M. S.; Jones, M.,

Eds.; Wiley-Interscience: Hoboken, NJ, 2004.

9

J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004 6373

A R T I C L E S

Ozerov et al.

Compound 32^H. We found that the hydrogenation of 13-

21 to alkane and 5 is slow (and suppressed by alkene) and that

the ¹H NMR spectrum of 13 exhibits only slight broadening of

some peaks in the presence of H₂(vide supra). These facts

suggest that the binding of H₂to 13-21 is weak, and this is

confirmed in the DFT results. The calculated structure of

(PNP^H)ReH(≡CMe)(H₂) (32^H) is reminiscent of the X-ray

structure³⁷of 13a*, with the dihydrogen ligand in place of C₂H₄.

The H-H distance of 0.88 Å is indicative of a very small

amount of back-donation into the σ* of H₂. The overall

geometry about Re is approximately octahedral. The Re-C and

Re-H(hydride) bond distances increase slightly compared to

the structure of 13^Hwhile the Re-N bond is much longer (0.15

Å) in 32^Hcompared to 13^Hdue to loss of N f Re π-donation.

The Re-N bond distance in 32^Hmatches that in 13a* quite

closely. We calculate that the dissociation of H₂from 32^His

exoergic (Scheme 15). This is consistent with the lack of

experimental observation of 32^Hat 22 °C.

Compound 22^H. We used DFT to investigate the possible

isomers of carbene compound 22^H. Since the observation of a

¹³C resonance at 273.1 ppm in the spectrum of 22b rules out

the olefin π-complex formulation (i.e., (PNP)ReH₂(η²-H₂Cd

CHOEt)) in favor of the carbene, we did not study the olefin

complex isomers by DFT methods. The preference for a carbene

(vs η²-olefin) complex will generally be greater for heteroatom-

substituted olefins.⁸Since the (PNP)ReH₂fragment prefers

the carbene isomer even for the cyclohexene/cyclohexylidene

system,⁵⁶we certainly expect unequivocal preference for a

carbene in the case of vinyl ethers. We calculated the mini-

mized geometries of the two isomers C (“central”) and S

(“side”) for 22^H. We find that the isomer S is lower in energy

by 6.7 kcal/mol. Given the fact that experimentally we observe

only one isomer of 22b and that the S isomer is also pre-

ferred⁵⁶for the related (PNP)Re(H)₂(dC(CH₂)₅), it is reason-

able to expect that the calculated preference for 22^H-S is

valid for 22b as well. The Re-C distance is considerably

shorter for 22^H-S as compared to that of 22^H-C. This indicates

that the back-donation into the empty p-orbital of the carbene

ligand is more effective when the carbene ligand is “on the side”

of two hydrides than when it is “in between” the two

hydrides. The diminished back-donation from Re in 22^H-C

is compensated by the increased π-donation from O as evi-

denced by the shorter C-O bond distance. Additionally, 22^H-S

possesses a short Re-H contact (2.610 Å) with one of the H’s

of the methoxy group. This weak agostic interaction may

contribute to the overall preference for 22^H-S. The DFT-

optimized structure of 22^H-S contains two classical hydrides.

Discounting the agostic Re-H interaction, the Re center has a

16-electron valence count. The agostic interaction competes for

the remaining empty orbital at Re with the p-lone pair of the

amido ligand. The Re-N bond in 22^H-C is shorter than that in

22^H-S.

(dC(CH₂)₅). The Re-N bond distance in 28^His intermediate

between those in (PNP)Re(H)₂(dC(CH₂)₅) and 22^H-S. For 28^H

it can also be argued that the π-donation from N is in

competition with the agostic interaction. The structure of

28^Hpossesses two classical hydrides. (PNP^H)Re(H)₂(η²-C₂H₄)

(25^H) and (PNP^H)ReH₃(CHdCH₂) (27^H) are isomeric to

28^H. We have obtained minimized energy geometries for two

isomers of 25^Hwhere the C₂H₄ligand is either between the

two hydrides (25^H-C) or on the side of them (25^H-S). In both

isomers, the C-C vector of the C₂H₄ligand is oriented

approximately parallel to the P-Re-P vector, i.e., perpen

dicular to the equatorial plane. Both structures are derived

from a pentagonal bipyramid, and for a Re^III(d⁴) center the

two filled d-orbitals are orthogonal to the equatorial plane, thus

the alignment of C₂H₄is dictated by the symmetry match

between the empty π* orbital of C₂H₄and the filled d-orbital

of Re. Notably, the placement of the carbene ligands of

22^Hand 28^Hin the equatorial plane is also dictated by the

symmetry of orbitals involved in back-donation. The Re-N

bond is shorter in 25^H-C than in 25^H-S, similar to the

corresponding isomers of 22^Hand 28^Heven though there are

no agostic contacts in either isomer of 25^H. This reaffirms that

the Re-N π-interaction is stronger for “C” isomers of com-

pounds of general formula (PNP)Re(H)₂(L) even when in the

“S” isomer there is no agostic CH donor competing with

π-donation from N.

The minimized structure of 27^His similar to that of 5, in

which one of the hydrides is replaced by a vinyl group.

However, in 27^Hthe distance between the two hydrogens

perpendicular to the PNP plane is shorter (1.349 Å) compared

to those in 5^H.

Compound 32^His also isomeric to 25^H, 27^H, and 28^H. The

relative free energies (kcal/mol) are presented in Scheme 16.

The most stable isomer is calculated to be 25^H-S, but, experi-

mentally, only 13, 13*, and 5 are observed as Re-containing

products when 5 is allowed to react with <4 equiv of C₂H₄.

The following reconciles the experimental observations with

the DFT results: formation of 13^Hand free H₂is endoergic by

only 2.2 kcal/mol with respect to 25^H-S and under experi-

mental conditions of low H₂concentration the equilibrium

will be shifted toward loss of H₂(and thus toward 13).

Formation of 13* will further stabilize the carbyne product

(ethylene binds to 13 more strongly than H₂). The model

PNP^Hligand is much less sterically demanding than the

experimentally used PNP^Rligands, and it should be expected

that the experimental system should favor smaller steric

profile ligands more so than the model DFT system. There-

fore, we expect the DFT model to overestimate the stability

of η²-olefin isomers relative to carbenes and hydrido car-

bynes

If we look only at the relative η²-ethylene/ethylidene

stability, we find that 25^H-S is more stable than 28^H-S by 11.1

kcal/mol. This difference is larger than what we found for the

cyclohexene/cyclohexylidene case and is perhaps too large

to be overridden by the steric effect of the experimental

PNP^Rligands. We find this consistent with the experimental

observation of an olefin isomer for (PNP^R)Re(H)₂(L) (L )

olefin or isomeric carbene) in the case of 29 and 30 (mono-

substituted olefins) and the carbene isomer for 32 (disubstituted

olefin).

Compounds 25^H, 27^H, and 28^H. The optimized geometry

of (PNP^H)Re(H)₂(dCHMe) (28^H) is similar to that of 22^H-S

and (PNP)Re(H)₂(dC(CH₂)₅). The ethylidene ligand shows an

angular distortion at the R-C with the methyl group of the

ethylidene bent toward Re, resulting in a close Re-H contact

(2.729 Å). A similar distortion was observed in (PNP)Re(H)₂-

(56) Ozerov, O. V.; Watson, L. A.; Pink, M.; Caulton, K. G. J. Am. Chem. Soc.

2003, 125, 9604.

9

6374 J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004

Transformation of Acyclic Alkenes to Hydrido Carbynes

A R T I C L E S

The spread of free energies for the isomers in Scheme 16 is

only 16.3 kcal/mol. Given some uncertainty in the DFT

numbers,⁵⁵particularly as discussed above, the existence of these

compounds cannot be ruled out as intermediates in the trans-

formation of alkenes to hydrido carbynes on the basis of their

thermodynamic stability.

the free energy of hydrogenation of ethylene, or any olefin, is

sufficient to reverse this situation and thus furnish a reactive

intermediate (e.g., [PNP]Re(H)₂(C₂H₄)). Although coordinated

olefin is more stable than coordinated carbene, this deficit is

overcome by elimination of H₂and formation of the hydrido

carbyne. Note also that H₂is not bound to [PNP]ReH(≡CMe),

consistent with the availability of H₂, in the reacting system, to

hydrogenate olefin. The ∆G°₂₉₈for H₂loss from [PNP]Re(H)-

(≡CMe)(H₂), -4.5 kcal/mol is dominated by a favorable entropy

(T∆S ) 9.3 kcal/mol at 298K), since the dissociation enthalpy

is only modest (+4.8 kcal/mol).⁵⁷

Isomers of 13^H. We have also examined several structures

isomeric to 13^Hto determine whether any obvious alternatives

may be thermodynamically more stable than 13^Hbut perhaps

kinetically inaccessible. The relative free energies are presented

in Scheme 17 (in kilocalories per mole). Clearly, 13^His the

thermodynamically preferred isomer. The coordination environ-

ment about Re in compounds 33^H, 34^H, and 35^His similar to

that in 25^H, 27^H, and 28^H, respectively, except for 35^H-S. 35^H-S

is better described (Figure 3) not as a vinylidene dihydride

but as a strongly R-agostic vinyl hydride, derived from the

former by partial migration of H from Re to C. 34^Hpossesses

an η²-H₂(d_HH) 0.987 Å) ligand trans to N. This H-H bond

shortens in the series of 5 f 27 f 34 as one of the hydrides in

5 is replaced by a vinyl (27) or ethynyl (33) group. Apparently

the more electron-withdrawing vinyl and ethynyl cause the Re

center to donate progressively less to the σ* of H₂. The

difference between the C and S isomers of 33^H(27.6 kcal/mol)

is much larger than that between the C and S isomers of 25^H

(3.4 kcal/mol). Unlike ethylene, acetylene can act as a four-

electron donor, but it can only do so in the S isomer. This may

account for the much larger energy difference. In addition,

acetylene, as a stronger π-acid than ethylene, may be more

sensitive to the degree of π-back-donation from Re in different

isomeric structures. Indeed, the acetylene C-C bond is 0.056

Å longer in 33^H-S than that in 33^H-C, while for the two isomers

of 25^Hthe difference in the ethylene C-C bond distances is

only 0.012 Å. For both 33^Hand 35^H, the Re-N bond is

substantially shorter in the C isomer in accord with the above

discussion.

Whereas the isomers in Scheme 16 consist of the elements

of ethane plus the (PNP^H)Re fragment, the isomers in Scheme

17 consist of the elements of ethylene and the (PNP^H)Re

fragment. Another possible isomer exists for alkenes higher than

ethylene: the allyl hydride structure 36^H. To assess its ther-

modynamic stability, we have calculated the free energy of a

hypothetical reaction (Scheme 18) between 13^Hand propene.

From that we found that 36^His slightly endoergic with respect

to 13^H. The preference for 13^His likely to be greater for the

experimental PNP^Rsystems and for alkenes higher than propene

because of steric factors.

Overall Reaction Free Energies. The reaction of 3 equiv

of C₂H₄with 5^Hto produce 13^Hand 2 equiv of C₂H₆is

calculated (Scheme 19) to be highly exoergic. However, our

calculations indicate that the transformation of ethylene into a

hydride and ethylidyne ligands is favorable even without

concomitant hydrogenation of sacrificial double bonds (Scheme

19). This finding supports our observation that under kinetic

control in the presence of excess alkene only 1 equiv of alkene

is converted to alkane and 1 equiv of free H₂is produced.

The free energy of binding H₂to [PNP]ReH₂is too large to

make eq 2 a viable

We have emphasized here how hydrogenation of added

ethylene by [PNP]Re(H)₄not only creates reactive intermediates,

but enhances ∆G°₂₉₈for the creation of such intermediates, as

well as of the final product. This is also illustrated as follows.

Equation 3 shows the standard free energy

[PNP]Re(H)₄+ C₂H₆^+30.08 [PNP]ReH(H₂)(tCMe) + 2H₂

(3)

C₂H₄+ H₂^-28.28 C₂H₆

(4)

for conversion of ethane to a carbyne complex to be unfavorable.

Only the addition of one or (better) two executions of

“sacrificial” olefin hydrogenation (eq 4) could enable alkane

conversion to alkylidyne.

We also examined the formation of 13^Hfrom methyl vinyl

ether (taken as a model for ethyl vinyl ether). We found

elimination of MeOH from 22^Hto produce 13^Hto be preferred

over the 1,2-shift of MeO to produce 37^H(Scheme 20).

Elimination of MeOH from 22^Hmight also produce compounds

isomeric to 13^H, but we have already shown here the thermo-

dynamic preference for 13^H. Both ROH elimination and RO

migration to the metal have been observed in related alkoxy-

carbene complexes of Os.^2,5

Conclusion

[PNP]Re(H)₄reacts with acyclic olefins under quite mild

conditions, with hydrogen removal from Re, to leave behind

lower valent reactive transients which act to strip H from

additional olefin to convert it to hydride and carbyne ligands.

[PNP]Re(H)₄^+15.08 [PNP]ReH₂+ H₂

mechanistic step at the temperatures employed here. However,

(2)

(57) Watson, L. A.; Eisenstein, O. J. Chem. Educ. 2002, 79, 1269.

9

J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004 6375

A R T I C L E S

Ozerov et al.

Table 3. Calculated Distances for Figure 3

Re−N

Re−C1

Re−C2

Re−C3

Re−H1

1.639 1.664 1.664 1.645

1.673 1.673

Re−H2

Re−H3

Re−H4

Re−PH

Re−O

H1−H2

H2−H3

H3−H4

C−O

C1−C2

C2−C3

3

5^H

2.062

2.345

2.210

-

1.525

-

7^H

-

1.694

-

10^H

13^H

-

1.665 1.668 1.791 1.791 2.347

1.607

-

0.921

-

2.123 1.756

-

1.649

1.637 1.639

1.627 1.654 2.610

1.651 1.658

1.630 1.651

1.650 1.667 1.667

1.633 1.652 2.729

1.681 1.822 1.822

1.680 1.689

1.702 1.657

1.669 1.732 1.732

1.638 1.643

1.657 1.978

-

1.469

22^H-C 2.066 2.059

-

1.341 1.519

1.354 1.503

22^H-S

2.123 1.987

3.046

-

25^H-C 2.021 2.309 2.309

-

1.414

1.426

1.336

1.514

2.117

1.259

1.315

1.226

1.312

1.334

25^H-S

27^H

2.110 2.203 2.203

2.068 2.046

2.147 1.903 2.820

2.274 1.763

-

1.349

-

28^H-S

32^H

-

0.880

-

33^H-C 2.004 2.238 2.238

-

33^H-S

34^H

2.131 2.013 2.013

-

2.017 2.034

-

0.987

-

35^H-C 2.077 1.937

-

35^H-S

36^H

2.013 1.930

2.030 2.178 2.244 2.378

2.281 1.762

-

1.648

1.723 1.718

-

1.437 1.402

37^H

-

2.040 0.998

-

1.390 1.465

-

Me₂S (2.55 mL, 35 mmol) was injected into the solution of NaI in

AcOH, and it was briefly stirred with a glass rod. The resulting AcOH

solution was added all at once to the HCl solution of HReO₄.

Immediately, a copious amount of I₂appeared in the flask. The flask

was plugged with a rubber septum and stirred for 2 h at 22 °C. Then

the suspension was poured onto a fritted glass funnel and filtered. The

solids on the filter were washed with 2 × 5 mL of glacial AcOH, 3 ×

10 mL of 5% HCl, 10 mL of glacial AcOH, and 2 × 5 mL of Et₂O. It

was possible to wash away all of I₂by extraction with AcOH or Et₂O,

but this was unnecessary. The solids (mixture of 3 and I₂) were

transferred to a Schlenk flask and were dried at 60 °C and 0.1 Torr for

24 h until all I₂was sublimed off and then transferred into the glovebox.

Yield: 4.6 g (76%). 3 decomposed when left in the air for several

hours, as evidenced by blackening of the solid or when 3 came into

contact with water or wet organic solvent. The decomposition was

apparently insignificant in an acidic solution. According to the results

of the elemental analysis, a small fraction of Cl in 3 ended up being

replaced by I. However, we saw no evidence of incorporation of iodide

into 4 when 3 was used as a starting material. Elemental Analysis.

Found: C, 9.74; H, 2.26; Hal, 34.68. Calcd for C₄H₁₂Cl₃OReS₂: C,

11.10; H, 2.79; Hal, 24.57. Calcd for C₄H₁₂Cl_2.28I_0.72OReS₂: C, 9.63;

H, 2.43; Hal, 34.53.

The latter represent both greater formal electron donor number

with regard to the 18-electron rule, but also greater π-acid

character, which is more compatible with the high π-basicity

of the electron rich [PNP]Re fragment. Reaction rates are

predictably influenced by increasing substitution on the vinyl

carbons, as is the formation constant for olefin binding to the

product [PNP]ReH(≡CR). Heteroatoms F or OR on a vinyl

carbon are cleaved under mild conditions, with liberation of

HF or HOR; these can be scavenged to diminish degradation

of the PNP ligand Si-N bonds. DFT calculations support and

extend structural results from X-ray diffraction, but provide

uniquely valuable energetic information on potential intermedi-

ates in these multistep transformations. There is abundant

evidence from Re-N bond lengths, both experimental and

computational, that the amide N in the PNP ligand can serve

as a π-donor to stabilize otherwise unsaturated species. In

addition, an examination of Tables 2 and 3 shows that the Si-N

distance varies with the combined σ + π basicity at N.

Experimental Section

General Considerations. All manipulations were performed using

standard Schlenk techniques or in an argon-filled glovebox unless

otherwise noted. All hydrocarbon reagents and solvents, Et₂O, THF,

C₆D₆, C₇D₈, and C₆D₁₂were vacuum transferred or distilled from NaK/

benzophenone/18-crown-6. PMe₃and Me₃SiCl were vacuum-transferred

and stored in the glovebox. Hexamethyldisiloxane was distilled from

NaK alloy. Ethyl vinyl ether was distilled from CaH₂. Other reagents

were used as received. The preparation of vinyl phenoxide in our labs

has been described.⁵Compounds 1a-c, 2c, 5a, 13a, and 13a* were

prepared as reported elsewhere.^{37 1}H NMR chemical shifts are reported

in ppm relative to protio impurities in the deuterated solvents. ³¹P spectra

are referenced to external standards of 85% H₃PO₄(at 0 ppm). NMR

(Me₂S)₂ReOBr₃(3-Br). 3-Br was prepared in a procedure identical

to that of 3, except using HBr instead of HCl. A yield of 0.91 g (57%)

of yellowish-green 3-Br was obtained starting from 1.0 g of HReO₄

(54 wt % Re, 2.8 mmol Re), 5 mL of concentrated HBr, NaI (0.90 g,

6.0 mmol), Me₂S (0.51 mL, 7.0 mmol), and 8 mL of AcOH.

(PNP^iPr)ReOCl₂(4b). (PNP^iPr)Li (1b) (3.00 g, 7.10 mmol) was

dissolved in 15 mL of THF, and MgCl₂(0.86 g, 9.0 mmol) was added

to it. This was vigorously stirred overnight, during which time the

mixture became homogeneous. The resulting solution was poured into

a flask containing a stirred suspension of (Me₂S)₂ReOCl₃(3) in 50 mL

of toluene/5 mL of 1,4-dioxane. The reaction mixture was stirred for

1 h, and then the volatiles were removed in vacuo. The residue was

treated with 20 mL of heptane, stirred for 5 min, and then the volatiles

were stripped again. The residue was extracted with pentane until the

washings were near colorless. The deep green filtrate was stripped to

ca. 20 mL and placed in the freezer at -30 °C for 24 h. The dark

green, somewhat glutinous precipitate was separated from the super-

natant by decantation, washed with cold pentane, and dried in vacuo

to give 3.20 g (68%) of 4b as a mixture of two isomers. Another 0.60

g (13%) of 4b was obtained by an analogous procedure from the

combined supernatant and pentane washings. Total yield: 3.80 g (81%).

1

spectra were recorded with a Varian Gemini 2000 (300 MHz H; 121

MHz ³¹P; 75 MHz ¹³C), a Varian Unity Inova instrument (400 MHz

¹H; 162 MHz ³¹P; 101 MHz ¹³C), or a Varian Unity Inova instrument

(500 MHz ¹H, 126 MHz ¹³C). Usage of cyclohexane-d₁₂or protio

solvents for NMR studies involving reactions of 5 with olefinic

substrates was necessitated by the fact that noticeable incorporation of

D into products occurred if C₆D₆was used as solvent. Elemental

analyses were performed by CALI, Inc. (Parsippany, NJ).

(Me₂S)₂ReOCl₃(3). No precautions against introduction of air were

taken during the synthesis of 3. HReO₄solution (5.0 g, 54 wt % Re,

14 mmol Re, Sandvik Chemicals) was treated with 25 mL of

concentrated HCl and was stirred in a 200 mL flask. In a separate flask,

NaI (4.5 g, 30 mmol) was dissolved in 50 mL of glacial acetic acid.

The mer,cis-Isomer (Major). H NMR (C₆D₆): δ 2.88 (m, J_HH

)

7 Hz, 4H, PCH), 2.20 (m, J_HH) 7 Hz, 4H, PCH), 1.51 (apparent q

(dvt), 8 Hz, 12H, PCH-CH₃), 1.20 (dvt, J_HH) 14 Hz, J_HP) 4 Hz,

9

6376 J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004

Transformation of Acyclic Alkenes to Hydrido Carbynes

A R T I C L E S

2H, PCH₂Si), 1.18 (apparent q (dvt), 8 Hz, 12H, PCH-CH₃), 1.16

(apparent q (dvt), 8 Hz, 12H, PCH-CH₃), 0.63 (dvt, J_HH) 14 Hz, J_HP

) 4 Hz, 2H, PCH₂Si), 0.56 (s, 6H, SiCH₃), 0.22 (s, 6H, SiCH₃). ³¹P-

{¹H} NMR (C₆D₆): δ 8.4 (s). ¹³C{¹H} NMR (C₆D₆): δ 27.6 (t, 14

Hz, P-CH), 25.7 (t, 11 Hz, P-CH), 18.8 (s, PCH-CH₃), 18.32 (s,

PCH-CH₃), 18.27 (s, PCH-CH₃), 18.2 (s, PCH-CH₃), 12.2 (t, 5 Hz,

PCH₂Si), 5.1 (t, 2 Hz, SiCH₃), 4.6 (br s, SiCH₃).

The flask used for the hydrogenation can be rinsed several times

with THF until only Mg powder remains, and then the volatiles were

removed in vacuo. The remaining Mg can be used repetitively for these

experiments, even for complexes of different PNP ligands.

(PNP^iPr)Re(H)₄(5b). Isolated yield of 2.09 g (65%) from 3.70 g

1

(5.56 mmol) of 4b. H NMR (C₆D₆): δ 1.90 (m, J_HH) 7 Hz, 4H,

PCH), 1.16 (apparent q (dvt), 8 Hz, 12H, PCH-CH₃), 0.98 (apparent

q (dvt), 8 Hz, 12H, PCH-CH₃), 0.75 (t, J_HP) 4 Hz, 4H, PCH₂Si),

0.32 (s, 12H, SiCH₃), -9.42 (t, 22 Hz, 4H, ReH). ³¹P{¹H} NMR

(C₆D₆): δ 66.5 (s). ¹³C{¹H} NMR (C₆D₆): δ 31.5 (t, 14 Hz, P-CH),

19.2 (t, 2 Hz, PCH-CH₃), 18.3 (s, PCH-CH₃), 9.8 (s, PCH₂Si), 6.5

(t, 3 Hz, SiCH₃).

1

The mer,trans-Isomer (Minor). H NMR (C₆D₆): δ 2.69 (m, J_HH

) 7 Hz, 4H, PCH), 1.41 (t, J_HP) 4 Hz, 4H, PCH₂Si), 1.31 (apparent

q (dvt), 8 Hz, 12H, PCH-CH₃), 1.20 (apparent q (dvt), 8 Hz, 12H,

PCH-CH₃), 0.22 (s, 12H, SiCH₃). ³¹P{¹H} NMR (C₆D₆): δ 16.7 (s).

¹³C{¹H} NMR (C₆D₆): δ 26.7 (t, 11 Hz, P-CH), 19.0 (s, PCH-CH₃),

18.7 (s, PCH-CH₃), 11.9 (t, 4 Hz, PCH₂Si), 5.2 (t, 2 Hz, SiCH₃).

(PNP^tBu)ReOCl₂(4c). (PNP^tBu)MgCl(dioxane) (2c) (3.38 g, 5.67

mmol) was added to a flask containing a stirred suspension of (Me₂S)₂-

ReOCl₃(3) in 30 mL of benzene/5 mL of 1,4-dioxane. The reaction

mixture was stirred for 1 h, and then the volatiles were removed in

vacuo. The workup essentially identical to that of 4b resulted in the

combined yield (two fractions) of 3.08 g (75%). Only the mer,cis-

isomer could be detected by NMR.

(PNP^tBu)Re(H)₄(5c). Isolated yield of 2.02 g (74%) from 3.08 g

1

(4.30 mmol) of 4c. H NMR (C₆D₆): δ 1.29 (vt, 7 Hz, 36H, CMe₃),

0.98 (t, J_HP) 5 Hz, 4H, PCH₂Si), 0.35 (s, 12H, SiCH₃), -9.20 (t, 22

Hz, 4H, ReH). ³¹P{¹H} NMR (C₆D₆): δ 83.6 (s). ¹³C{¹H} NMR

(C₆D₆): δ 37.8 (t, 10 Hz, P-CMe₃), 30.1 (s, PCMe₃), 10.2 (s, PCH₂-

Si), 6.4 (t, 2 Hz, SiCH₃). Elemental Analysis. Calcd (Found) for C₂₂H₅₆-

NP₂ReSi₂: C, 41.35 (41.40); H, 8.83 (9.48); N, 2.19 (2.18).

(PNP^Cy)ReO(H)₂(7a). (PNP^Cy)ReOCl₂(4a, 0.322 g, 0.390 mmol)

was dissolved in 10 mL of C₆H₆. To this solution NaBEt₃H in toluene

(0.780 mL of 1.0 M solution, 0.780 mmol) was added. The color rapidly

changed from green to brown. The reaction was stirred for 2 h at

ambient temperature, and then the volatiles were removed in vacuo.

The residue was treated with heptane, and the volatiles were stripped

again. The residue was extracted with toluene and filtered. The filtrate

was stripped, and the solids were washed with pentane and dried in

vacuo to give 0.24 g (81%) of yellow 7a.

¹H NMR (C₆D₆): δ 1.46 (vt, 7 Hz, 18H, CMe₃), 1.34 (dvt, J_HH

)

14 Hz, J_HP) 4 Hz, 2H, PCH₂Si), 1.24 (vt, 7 Hz, 18H, CMe₃), 0.61

(dvt, J_HH) 14 Hz, J_HP) 4 Hz, 2H, PCH₂Si), 0.59 (s, 6H, SiCH₃),

0.27 (s, 6H, SiCH₃). ³¹P{¹H} NMR (C₆D₆): δ 6.7 (s).

(POP^iPr)ReNCl₂(6b). A sample of 4b was heated in C₆D₆for 5 h

at 80 °C. Complete disappearance of 4b was observed. The major

component of the reaction mixture (>85%) was 6b.

1

The mer,trans-Isomer (Major). H NMR (C₆D₆): δ 2.94 (m, J_HH

¹H NMR (C₆D₆): δ 5.59 (t, 16 Hz, 2H, ReH), 2.22 (br t, 8H), 1.5-

2.0 (several multiplets, 28H), 1.1-1.35 (several multiplets, 12H), 0.25

(s, 12H, SiCH₃). ³¹P{¹H} NMR (C₆D₆): δ 35.6 (s). ¹³C{¹H} NMR

(C₆D₆): δ 37.7 (t, 12 Hz, P-CH), 29.5 (s, CH₂of Cy), 29.3 (s, CH₂of

Cy), 27.5 (t, 5 Hz, CH₂of Cy), 27.4 (t, 5 Hz, CH₂of Cy), 26.7 (s, CH₂

of Cy), 20.1 (t, 4 Hz PCH₂Si), 3.1 (s, SiCH₃). Elemental Analysis.

Calcd (Found) for C₃₀H₆₂NOP₂ReSi₂: C, 47.59 (47.59); H, 8.25 (8.48);

N, 1.85 (1.58).

(PNP^Cy)Re(H)₄(PMe₃) (10a). 5a (0.233 g, 0.300 mmol) was

dissolved in 3 mL of pentane, and ca. 0.1 mL of PMe₃was added to

this solution. The solution became essentially colorless in the time of

mixing. The solution was placed into a -30 °C freezer overnight. The

next day the colorless crystalline (X-ray quality crystals) 10a was

isolated by decantation of the supernatant, washed with cold pentane,

and blow-dried using a glass pipet inside the glovebox. Yield: 0.21 g

(86%).

¹H NMR (C₇D₈, 20 °C): δ 2.12 (br s, 2H), 1.88 (br s, 12 H), 1.69

(br s, 12H), 1.42 (br s, 4H), 1.43 (dt, 8 Hz, 1 Hz, PMe₃), 1.27 (br m,

14 H), 0.88 (br t, 5 Hz, PCH₂Si), 0.39 (s, 12 H, SiMe). At 20 °C, the

hydride signals were extremely broad and they could not be observed.

¹H NMR (C₇D₈, -76 °C, selected resonances, fine structure not

completely resolved): δ -1.38 (br t, 28 Hz, 1H, Re-H), -2.20 (br dd,

28 Hz, 19 Hz, 2H, Re-H), -9.30 (br d, 72 Hz, 1H, Re-H). ³¹P{¹H}

NMR (C₆D₆, 22 °C): δ 33.1 (d, 15 Hz, 2P, PCy₂), -33.6 (br t, 1P,

PMe₃). ³¹P{¹H} NMR (C₇D₈, -76 °C): δ 33.5 (d, 15 Hz, 2P, PCy₂),

-32.4 (t, 15 Hz, 1P, PMe₃). ¹³C{¹H} NMR (C₇D₈, 20 °C): δ 37.8 (t,

10 Hz, P-CH), 30.7 (d, 36 Hz, PMe₃), 29.4 (s, CH₂of Cy), 28.1 (s,

CH₂of Cy), 27.7 (two triplets overlapping, CH₂of Cy), 27.2 (s, CH₂

of Cy), 17.3 (br s, PCH₂Si), 4.6 (s, SiCH₃).

(PNP^Cy)Re(H)₂(PMe₃) (11a). 5a (0.467 g, 0.600 mmol) was

dissolved in 10 mL of pentane, and ca. 0.1 mL of PMe₃was added to

this solution. The solution became essentially colorless in the time of

mixing. Stripping the volatiles in vacuo produced mostly the colorless

10a with some purple 11a. This residue was heated at 120 °C under

0.1 Torr for 1 h and then redissolved in 5 mL of toluene, and the cycle

was repeated twice more. NMR indicated complete conversion to 11a.

Analytically pure 11a can be obtained by recrystallization from pentane

at -30 °C. Yield: 0.410 g (84%).

) 7 Hz, 4H, PCH), 1.51 (apparent q (dvt), 8 Hz, 12H, PCH-CH₃),

1.28 (t, J_HP) 4 Hz, 4H, PCH₂Si), 1.19 (apparent q (dvt), 8 Hz, 12H,

PCH-CH₃), 0.09 (s, 12H, SiCH₃). ³¹P{¹H} NMR (C₆D₆): δ 24.9 (s).

¹³C{¹H} NMR (C₆D₆): δ 23.3 (t, 12 Hz, P-CH), 19.9 (s, PCH-CH₃),

18.3 (s, PCH-CH₃), 12.1 (t, 6 Hz, PCH₂Si), 2.7 (s, SiCH₃).

(POP^tBu)ReNCl₂(6c). A sample of 4c was heated in C₆D₆for 1 h

at 80 °C. Full conversion to 6c (mer,trans-isomer only) was observed

by NMR.

¹H NMR (C₆D₆): δ 1.51 (vt, 7 Hz, 36H, CMe₃), 1.24 (vt, J_HP) 4

Hz, 4H, PCH₂Si), 0.24 (s, 12H, SiCH₃). ³¹P{¹H} NMR (C₆D₆): δ 38.4

(s).

Syntheses of 5a-c. The Mg powder to be used was activated in a

separate Schlenk flask by being stirred in a glovebox overnight in THF

with Me₃SiCl (Si/Mg 0.02), and then the Mg powder was rinsed several

times with THF to remove nonmetallic solids and dried in vacuo. A

large excess (∼3 g) of Mg thus activated and a stir bar were placed

into a 300-mL flask equipped with a Teflon vacuum valve and a

SolvSeal (Andrews Glass) capped joint. In an Ar-filled glovebox

(without stirring!) a precursor 4a-c was added, along with 25-50 mL

of Et₂O and ca. 20% MgBr₂(Et₂O) (molar, per Re). The flask was taken

out of the glovebox, and the contents were degassed by a freeze-

pump-thaw cycle. Then the flask was allowed to warm to ca. -20

°C, back-filled with ∼ 1 atm H₂using a gas line equipped with a

manometer, and closed off. At this point vigorous stirring was initiated.

With Mg activated as described here, the intense color of 5a-c starts

to appear within 15 min. H₂was replenished at intervals. The reaction

mixture may become slightly warm. The stirring was continued

overnight, after which the volatiles were stripped to dryness. The residue

was treated with heptane/dioxane and stripped, and then treated with

heptane and stripped. The residue was then extracted with pentane and

filtered, and the filtrate was stripped to dryness. This residue was

dissolved in an appropriate solvent and filtered, and the product was

obtained by crystallization at -30 °C, in two or more fractions, if

necessary. 5a was best crystallized out of pentane or a mixture of

pentane with another alkane. 5b and 5c were best crystallized out of

mixtures of neo-hexane with Me₄Si or Me₃SiOSiMe₃. The precipitates

of 5b and 5c, after decantation of the supernatant, should be washed

with cold Me₄Si followed by immediate drying in vacuo.

9

J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004 6377

A R T I C L E S

Ozerov et al.

¹H NMR (C₆D₆): δ 2.38 (br d, 48H), 1.75-1.88 (br m, 12H), 1.60-

1.73 (multiplets, 8H), 1.58 (d, 6 Hz, 9H, PMe₃), 1.12-1.46 (multiplets,

20H), 0.79 (br t, 4H, PCH₂Si), 0.46 (s, 12H, SiCH₃), -11.24 (dt, 75

Hz, 11 Hz, 2H, ReH). ³¹P{¹H} NMR (C₆D₆): δ 32.5 (d, 14 Hz, 2P,

PCy₂), -66.5 (t, 14 Hz, 1P, PMe₃). ³¹P NMR with selective decoupling

of alkyl hydrogens showed a dt at 32.5 ppm and a tt at -66.5 ppm.

¹³C{¹H} NMR (C₆D₆): δ 41.2 (t, 10 Hz, P-CH), 34.1 (d, 27 Hz, PMe₃),

30.3 (s, CH₂of Cy), 29.3 (s, CH₂of Cy), 27.9 (t, 4 Hz, CH₂of Cy),

27.7 (t, 6 Hz, CH₂of Cy), 27.2 (s, CH₂of Cy), 13.0 (d, 4 Hz, PCH₂-

Si), 7.2 (s, SiCH₃). Elemental Analysis. Calcd (Found) for C₃₃H₇₁NP₃-

ReSi₂: C, 48.50 (48.24); H, 8.76 (8.95); N, 1.71 (1.63).

(PNP^iPr)ReH(≡C-CH₃) (13b) and (PNP^iPr)ReH(≡C-CH₃)(C₂H₄)

(13*b). (PNP^iPr)ReH₄(70 mg, 120 µmol) was dissolved in 2 mL of

heptane in a Teflon-stoppered 50-mL glass vessel. The solution was

degassed, and the vessel was filled with 1 atm of ethylene. The reaction

yielded an essentially colorless solution in under 5 min. The volatiles

were removed in vacuo at ambient temperature. The residue was

dissolved in C₆D₆, and NMR indicated an essentially quantitative

conversion to (PNP^iPr)ReH(C-CH₃)(C₂H₄). The volatiles were removed

in vacuo and the off-white residue was dried in vacuo at 160 °C for 30

min, during which time the color changed to the red of (PNP^iPr)ReH-

(C-CH₃). This red oil was dissolved in Me₃SiOSiMe₃(1 mL) and

filtered through a pad of Celite to remove a minor amount of insoluble

material. The filtrate was placed into a -30 °C freezer, and after 24 h

the red crystalline product was separated by decantation and dried in

vacuo. Yield: 45 mg (60%).

Experimental details for 13c-21b* and 22b-30b are given as

Supporting Information.

(PNP^iPr)ReH(≡C-CH₃)(H₂CdCH-OPh) (13b*OPh). 5b (30 mg,

52 µmol) was dissolved in 0.5 mL of protio-cyclohexane, and 0.1 mL

of Me₂NSiMe₃was added to this solution. No change was detected by

³¹P NMR when this solution was allowed to stand for 1 day at ambient

temperature. Vinyl phenoxide (19 µL, ca. 155 µmol) was added to this

solution. The solution became colorless within minutes, and solids

precipitated. Toluene (0.3 mL) was added to dissolve the precipitate,

and a homogeneous solution resulted. ³¹P NMR showed the presence

of the two isomers of 13b*OPh. The spectrum collected after this

solution was allowed to stand for 10 days at ambient temperature

showed no change. Removing the volatiles from this solution in vacuo

results in formation of a mixture of 13b and 13b*OPh, but quantitative

removal of the complexed vinyl phenoxide is difficult because of the

low volatility of the free vinyl phenoxide. The two isomers of 13b*OPh

are exclusively observed (³¹P NMR) when a solution of 13b is treated

with vinyl phenoxide.

Major isomer ³¹P{¹H} NMR (toluene/cyclohexane/Me₂NSiMe₃): δ

46.5 (d, 110 Hz), 42.8 (d, 110 Hz).

Minor isomer ³¹P{¹H} NMR (toluene/cyclohexane/Me₂NSiMe₃): δ

44.9 (d, 107 Hz), 43.2 (d, 107 Hz).

(PNP^iPr)ReH(≡C-CH₃)(H₂CdCH-F) (13b*F). 5b (30 mg, 52

µmol) was dissolved in 0.5 mL of protio-cyclohexane, and 0.1 mL

Me₂NSiMe₃was added to this solution. The solution was degassed,

and the NMR tube was back-filled with ca. 1 atm of vinyl fluoride.

Within the time of mixing, the color changed to yellow-brown. ³¹P

NMR analysis showed that the major isomer of 13b*F was the major

(>85%) component of the mixture. ¹⁹F NMR indicated the presence

of a significant amount of Me₃SiF (-158.4 ppm). A few other minor

peaks were observed. The volatiles were removed in vacuo, and the

residue was dried at 100 °C and 0.1 Torr. The red color of 13b

developed, and it was confirmed as the predominant component of the

1

(PNP^iPr)ReH(≡C-CH₃) (13b). H NMR (C₆D₆): δ 2.09 (m, J_HH

) 7 Hz, 2H, PCH), 1.83 (m, J_HH) 7 Hz, 2H, PCH), 1.23 (apparent q

(dvt), 8 Hz, 6H, PCH-CH₃), 1.15 (apparent q (dvt), 8 Hz, 6H, PCH-

CH₃), 1.14 (t, 3H, Re≡C-CH₃), 1.07 (apparent q (dvt), 8 Hz, 6H,

PCH-CH₃), 1.04 (apparent q (dvt), 8 Hz, 6H, PCH-CH₃), 0.85 (AB

dvt, J_HH) 14 Hz, J_HP) 4 Hz, 2H, PCH₂Si), 0.73 (AB dvt, J_HH) 14

Hz, J_HP) 4 Hz, 2H, PCH₂Si), 0.35 (s, 6H, SiCH₃), 0.30 (s, 6H, SiCH₃),

-10.15 (t, 14 Hz, 1H, ReH). ³¹P{¹H} NMR (C₆D₆): δ 61.09 (s). ¹³C-

{¹H} NMR (C₆D₆): δ 271.5 (t, 11 Hz, Re≡C), 38.1 (s, Re≡C-CH₃),

29.6 (t, 12 Hz, P-CH), 29.0 (t, 11 Hz, P-CH), 20.0 (br s, PCH-

CH₃), 19.0 (s, PCH-CH₃), 18.6 (s, PCH-CH₃), 17.6 (s, PCH-CH₃),

10.5 (s, PCH₂Si), 7.3 (s, SiCH₃), 6.3 (t, 2 Hz, SiCH₃).

1

mixture by ³¹P, H, and ¹³C NMR (C₆D₆solution). When solutions of

13b were exposed to the atmosphere of C₂H₃F, only two ¹⁹F peaks

were observed. Both were also observed in the in situ reaction mixture

above and we assigned them as the major and minor isomers of 13b*F.

13b*F in protio-cyclohexane/Me₂NSiMe₃. Major isomer. ³¹P{¹H}

NMR: δ 44.7 (d, 111 Hz), 42.4 (d, 111 Hz). ¹⁹F NMR: δ -170.8

(apparent dt, 68 Hz, 23 Hz). ¹H NMR (selected resonances, cyclohexane

peak set to 1.40 ppm): δ 6.02 (ddd, J_HF) 68 Hz, Re(H₂CdCHF)).

1

(PNP^iPr)ReH(≡C-CH₃)(C₂H₄) (13*b). H NMR (C₆D₆): δ 2.39

(br d, 10 Hz, 2H, C₂H₄), 2.23 (m, J_HH) 7 Hz, 2H, PCH), 2.21 (t, 32

Hz, 1H, ReH), 2.13 (m, J_HH) 7 Hz, 2H, PCH), 1.48 (br m, 2H, C₂H₄),

1.31 (AB dvt, J_HH) 14 Hz, J_HP) 4 Hz, 2H, PCH₂Si), 1.30 (apparent

q (dvt), 8 Hz, 6H, PCH-CH₃), 1.13-1.17 (three apparent q (dvt)

overlapping, 18H, PCH-CH₃), 1.01 (t, 3H, Re≡C-CH₃), 0.98 (AB

dvt, J_HH) 14 Hz, J_HP) 4 Hz, 2H, PCH₂Si), 0.21 (s, 6H, SiCH₃), 0.17

Minor Isomer. ¹⁹F NMR: δ -195.9 (doublet of multiplets, J_HF

67 Hz).

)

Acknowledgment. This work was supported by the National

Science Foundation.

1

(s, 6H, SiCH₃). H{³¹P} NMR (C₆D₆, selected resonances): δ 2.39

Supporting Information Available: Full crystallographic

narrative descriptions and data (CIF), computational methods

and DFT calculated structures, together with experimental

synthesis and spectra of compounds not in the main text (PDF).

This material is available free of charge via the Internet at

http://pubs.acs.org.

(approximately d, 10 Hz, 2H, C₂H₄), 1.48 (approximately d, 10 Hz,

2H, C₂H₄), peaks at 2.39 and 1.48 ppm constitute an AA′XX′ system.

³¹P{¹H} NMR (C₆D₆): δ 44.9 (s). ¹³C{¹H} NMR (C₆D₆): δ 259.6 (t,

11 Hz, Re≡C), 37.8(s, Re≡C-CH₃), 32.7 (t, 11 Hz, P-CH), 30.4 (br

s, C₂H₄), 29.3 (t, 15 Hz, P-CH), 20.6 (s, PCH-CH₃), 20.3 (s, PCH-

CH₃), 20.2 (t, 5 Hz, PCH₂Si), 19.9 (s, PCH-CH₃), 18.8 (s, PCH-

CH₃), 8.3 (t, 3 Hz, SiCH₃), 5.8 (br s, SiCH₃).

JA031617U

9

6378 J. AM. CHEM. SOC. VOL. 126, NO. 20, 2004

Article Doi

Transformation of acyclic alkenes to hydrido carbynes by (PNP R)Re complexes

DOI: 10.1021/ja031617u

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja031617u

Full text of DOI:10.1021/ja031617u

Products guided by the article

R&D Labs maybe for 714264-19-4

Relevant to this article

Hot Product